Until now, Sony had a solid track record of implementing artificial intelligence (AI) into its cameras. Starting with Real-Time Eye Autofocus (AF) in the Sony A7R II, it laid the groundwork for modern AI subject recognition in its latest cameras such as the A7R VI.

However, the latest AI photo assistant feature that has been rolled out in the new Sony Xperia 1 VIII camera phone feels like another spoonful of unnecessary AI cough medicine being force-fed to us, just like every other tech company on Earth seems to be doing since the AI gold rush began a few years ago.

Essentially, the new Xperia 1 VIII AI assistant suggests four different shooting settings at moment you capture the shot, each offering a distinct creative aesthetic based on the scene or subject.

According to Sony, the assistant helps you “discover new ways of shooting” all while “expanding the joy of creating”. Jesus, I’m cringing.

ABOVE: the new Sony Xperia 1 VIII and its AI assistant explained

As you can expect, this hasn’t resonated well with photographers, who've made it clear in forum chatter and publications that this AI assistant is nothing but a waste of battery life dressed up as a creative outlet.

And I couldn’t agree more. What a load of baloney about it helping you to “discover new ways of shooting”. The AI assistant essentially adds post-production edits only during production, which hardly broadens your compositional horizons.

Having the AI analyze the scene before you press the shutter and suggest compositional changes would be impressive. I'd prefer to come to the conclusion myself without the help of AI but, admittedly, a feature like this could help you explore new angles.

But the Xperia 1 VIII AI assistant is far from this and, at best, its achievement has been compressing production and post-production into a single click of the shutter.

Was this necessary? Perhaps, if you’re in the game of saving time when taking photos. But for anyone who actually enjoys the traditional photographic process, it brings nothing to the table.

I’d like to think that Sony’s new AI assistant could help photographers develop their editorial eye by explaining what edits it actually made when coming up with its suggestions. However, as it simply creates these at the click of a button, you don’t even get the educational benefit.

To me, it’s simply creative outsourcing to a machine and the complete opposite of “expanding your range of expression” (another quote that came directly from the mouth of Sony).

As a longtime Sony shooter, I’m disappointed by this new AI assistant and can feel my gag reflex kicking in as I regurgitate the thought of this unnecessary feature.

I’m really hoping Sony doesn’t have the bright idea of bringing it to Alpha cameras, as I’m currently looking to upgrade from the A7 III to a newer model – and I have a slight suspicion that a future firmware update could see my next camera infected by this trash.

You might also like…

Unlike the Xperia 1 VIII smartphone, I am a big fan of the company's cameras – and we have an expertly-curated list of the best Sony cameras here.

AI this, AI that – I can’t scroll through anything on the internet without being inundated with generative AI. As both a photographer and a writer – two art forms that generative AI attempts to imitate – I’m AI-exhausted. But here’s the thing – I’m not the only one, and the latest data seems to suggest around half the population is just as frustrated with AI as I am.

Around half of all adults in the US are wary of AI’s impact, according to a June 2025 PEW survey – a sharp increase from the 37 percent of AI-weary humans in 2021.

A new YouGov survey suggests that British adults share a similar stance, with 42 percent saying that they would get rid of generative AI entirely and 22 percent not even sure if they like the technology.

Only around 18.3 percent of US adults call themselves artists (and only 1.6 percent professional artists), which suggests that the artists whose jobs are most threatened by generative AI are far from being the only humans wary of the technology.

My stance on AI comes as someone who has tried out bits and pieces of the technology. As a journalist, it’s my job to explore the pros and cons of generative AI in order to better inform my readers, so I’ve done everything from asking ChatGPT to generate images that still look like me to testing out Photoshop’s new AI assistant.

Yet, even after all this testing, I’ve turned off Apple Intelligence on my iPhone, I leave generative AI off on most Photoshop tools where possible, and I continue to use the camera and the keyboard to create.

The problem is – and probably why so many adults agree with me – AI feels impossible to turn off. I can’t keep it out of my social media feeds, and Google doesn’t have an off switch for its AI Overviews. I can ignore generative AI, but I can’t ignore the rising prices on memory cards and hard drives that it has created. Artists have just been told to accept that AI is the norm and move on – or get left behind.

To be clear, I don’t think all AI is bad. Eye-detection autofocus is one of my favorite features on modern mirrorless cameras, and, yes, that’s an AI technology. Lightroom’s auto mask features to select skin, teeth, or eyes in a portrait have saved me a lot of time. The ability to search through the photo library on my phone is one of my favorite iPhone updates.

But, while those tools are AI, they are not generative AI. I don’t mind an AI that’s going to help save me from tedious tasks. But I’m a creator because I love creating, and I don’t want a computer to do it for me.

I’m not against AI, I’m wary of generative AI, the type of AI that creates something from nothing. (Or, more accurately, that creates something from a library of work created by real human artists that, in many cases, was stolen without permission.)

Generative AI also tends to have a much larger impact on the environment than non-generative AI. Generative AI is the technology that requires massive data centers, creating noise nuisance in small communities and driving up the cost of any technology that requires digital storage.

But these views are perhaps unsurprising coming from someone who earns a living as a creator. What’s giving me hope is the number of non-artists who are just as frustrated by AI as I am.

Even Gen Z – the generation with the highest adoption of generative AI – is becoming more wary of the technology. A recent Gallup poll has found that the number of Gen Z-ers excited about AI has dropped while the number reporting anger has increased.

All of this is happening at the same time as some Big Tech companies are reporting that the cost of generative AI is higher than having human employees.

At the same time, “AI Slop” was the Word of the Year for 2025, hinting that many would prefer to see human-created content in their social media feeds.

All of the data on the increasing number of people wary of AI technology gives me hope that even non-artists are going to continue to value human-created art. I don’t think AI is going anywhere, but my hope is that the number of people wary of AI will temper its growth – and that even non-artists will continue to appreciate human artwork.

You may also like...

Browse the best cameras for beginners to take real photographs.

Last week, the Tribeca Film Festival announced that, at this year’s edition of the event, it will screen a documentary film about the struggles of Iranians living under the oppressive regime in their country.

But this isn't just another feature about struggle and survival amid political turmoil. The documentary is the first feature-length production entirely created using AI to be showcased at a major film festival.

Dreams of Violets is a 75-minute docudrama inspired by real events that happened earlier this year in protests in Tehran. Thanks to AI, was generated on a budget of just $2,000 (approximately £1,500 / AU$2,800), involving no actors or a single set being built.

Above: Dreams of Violets trailer

In a statement, director Ash Koosha admitted that he would have preferred to make the film with more traditional methods. But, as an Iranian in exile, the director said that AI was the only way that he could tell the story.

"The film exists because the dead deserve to be witnessed and because the families inside Iran, who cannot speak, deserve someone outside who refuses to forget," the director wrote. "I am not certain that this is the right form for this story. I am certain that this story needed a form, and this is the form that was available to me."

AI technology is making it possible to create works based in regions where travel is dangerous. But, the project's inclusion in the film festival is a controversial one.

The inclusion of a fully AI-generated film at a major film festival could mark a monumental turning point in filmmaking: one that sets off a steady slide toward widespread public acceptance of fully AI-generated movies and, subsequently, the vast majority of production companies cashing in on the budget-saving benefits of using the technology instead of real people and locations.

If you watch the Dreams of Violets trailer, you can clearly tell that the movie is an AI creation, with some characteristic jerky movements and the sometimes jarring stares of people who were conjured up by an algorithm.

However, to my eye, it isn't far off from passing for real live action, which is frightening to think about, considering that the gold rush to develop AI film generation platforms capable of creating realistic footage only began a couple of years ago, spearheaded by the release of tools such as OpenAI’s Sora in 2024 – which was recently shut down.

If, after just two or so years, AI is now advanced enough to create a documentary deemed worthy of screening at a major film festival, are we just a few years away from completely AI-generated films hitting the big screens in our local cinemas too?

Perhaps so. Only recently, Disney laid off some 1,000 staff members in what it claimed was a move toward a “culture of efficiency."

And, even more recently, esteemed director Martin Scorsese publicly embraced AI, becoming an advisor to an AI movie-generation tool startup.

Will we ever reach a point where mainstream cinema completely does away with real humans and locations appearing on screen? I don't think so.

For that to happen, AI tools still need a lot more refinement, and I think there will always be an audience and production companies that will only ever want to immerse themselves in human cinematography, regardless of whether AI films become the norm in a not-so-distant dystopian future.

But with Dreams of Violets already proving just how close AI tools are to replacing even the most skilled camera operators, actors, and editors, and the Tribeca Film Festival redefining what is accepted as cinematography, I’m absolutely certain that, as a 30-year-old writing this, fully generated AI movies could become the norm within my lifetime, more likely within the next couple of decades.

The 25th Tribeca Film Festival is set to run 03-14 June in New York City, with Dreams of Violets premiering on June 10.

You may also like...

A photo of an owl was dethroned after winning a wildlife competition having been AI generated. Is 'genAI' paranoia ruining photography contests?

Sometimes, an event feels so outrageous or unlikely, it's difficult as a journalist to resist the well-worn cliches we're trained to avoid: "You couldn't make it up!"; "It's like something out of Black Mirror!" and so on. Well, such an event has just taken place, and I have to be honest: I'm lost for words. So I'm just going to skip the formalities, and simply tell you what took place.

One of the most famous photographs ever taken, taken by a man who spent his life fighting for photography to be taken seriously as fine art, has now been AI-colorised, printed in editions of ten, and offered for sale at a prestigious New York gallery. Without anyone bothering to ask if that was okay.

As reported by Art News the Danziger Gallery, a well-regarded photography dealer in Manhattan, exhibited what it described as an "A.I. Generated" color version of Ansel Adams' Moonrise, Hernandez, New Mexico at the Association of International Photography Art Dealers' annual show, The Photography Show, at the Park Avenue Armory last month. The show ran from April 22-26 2026. Works at this event have previously sold for tens of thousands of dollars.

The image was created from the prompt "Make a realistic color version of Ansel Adams' iconic 'Moonrise Over Hernandez'" and was then, according to its own description, "proofed, regenerated, and photoshopped" over several months before being printed by master printer Esteban Mauchi in three sizes.

There's just one problem: nobody thought to tell the Ansel Adams Publishing Rights Trust it existed.

Why his legacy matters

Here's some context for the uninitiated. Ansel Adams (1902–1984) was a photographer and environmentalist whose large-format black-and-white landscapes, particularly of the American West, are among the most recognised images in photographic history. He pioneered the Zone System, a technical framework for controlling exposure and development to achieve a full tonal range, and co-founded the photography department at the Museum of Modern Art in New York.

More than perhaps any other photographer of his era, he dedicated his career to the argument that the camera was as legitimate an artistic tool as a brush or a chisel. Moonrise, Hernandez, New Mexico, shot in 1941, is his most famous single image, and originals have sold at auction for over half a million dollars.

Interestingly, the Trust's public statement, published on Instagram over the weekend, isn't quite the AI-bashing screed you might expect. They were at pains to point out that Adams himself would have been fascinated by AI tools. "Ansel was an innovator who expanded the expressive and technical possibilities of his medium," the statement reads. "He was remarkably prescient about, and excited by, the potential of computers to transform photography."

Their issue, then, isn't the use of AI itself. It's the use of creative content without permission. That's something I obviously feel strongly about myself as a writer, and I'm pretty sure most photographers do too. However James Danziger, the owner of the gallery, doesn't believe he's done anything wrong. “As the image is in the public domain I had every right to create a new and transformative work,” he wrote in a public statement.

Legally speaking, he may well be right. But morally, I'm not so sure. The Trust's statement noted that "few figures fought harder than Ansel to secure photography's place as fine art, or contributed more to the cultural conditions that gave rise to today's photography market." James Danziger talked of his "love of the iconic image"; it's a shame that love didn't extend to a courtesy call.

Adobe’s Agentic AI – the AI that runs tasks for you – has only just arrived on scene, but Adobe is already bringing the ability to use apps like Photoshop and Premiere into Google’s AI. During Google I/O, the two companies announced that Adobe apps will soon connect to Gemini.

The move will allow Gemini users to describe what they want while the agentic AI moves through the steps inside Adobe apps, including photo, video, and design software. Adobe says that users will also be able to continue working on the project inside the Adobe app itself or with Adobe Firefly.

The Gemini-Adobe integration will arrive “in the coming weeks” and Adobe hasn’t spelled out details on exactly what the agentic AI can and cannot do yet. But a few examples of how the integration has been used in existing agentic options offers hints that using Gemini will allow photographers to edit images without the back-and-forth between Lightroom and Photoshop.

Photographers can already use Firefly AI Assistant to edit photos across both Lightroom and Photoshop without the constant switch between the apps. As Adobe explains, one photographer using Firefly AI Assistant “let the assistant work through it step by step, connecting the right Adobe tools, checking in for feedback, and moving forward only when he approved.”

The example of getting feedback at every step is an important one. I’ve tested the agentic AI inside the beta version of Photoshop Web, and it felt a bit like an overenthusiastic photo intern – I didn’t always get the results I had asked for.

I’ve also tried ChatGPT’s Photoshop integration and was surprised at how the chatbot leaned towards a learning experience rather than an easy fix – and I’m hopeful that the ucpming Adobe-Gemini could similarly be used to learn the steps of more complex edits.

Another important distinction is that while Agentic AI is built on generative AI, Agentic AI can be used without generating a new image. The AI assistant in the beta Photoshop Web, for example, walks through the same steps in Photoshop that I would manually adjust – it doesn’t generate new pixels unless using a generative tool like Generative Fill.

Gemini will soon join Adobe’s own Firefly AI Assistant and Claude as AIs that are capable of carrying out steps inside Adobe apps, including Premiere, Illustrator and Express as well as Photoshop.

Adobe says that “we’re just getting started" with agentic AI. “There’s tremendous momentum around agentic creativity and we have a great roadmap that we’re bringing creators,” Adobe’s Forest Key, Vice President for Agentic AI for Creativity & Productivity Business, wrote in a blog post. “But what excites us most is what creators are already doing with these tools. The workflows you’re inventing. The ideas you’re bringing to life that simply weren't feasible before.”

Adobe hasn’t yet shared the cost of the tool. However, it’s based on the existing connector that works with Claude, which includes around 40 tools free, with additional tools for creators with Adobe subscriptions.

You may also like

Browse the best photo editing software or the best Photoshop alternatives.

Google’s Nano Banana brought an AI photo generator that could create images that resembled real people – and now Google is giving the same treatment to generated video. During Google I/O, the tech giant announced Gemini Omni, a new AI that allows users to generate video from existing videos, photos, or text.

Like Gemini 2.5 Flash Image – better known by its viral code name Nano Banana – Gemini Omni aims to keep the likeness of the person in the inspiration photo or video intact, a task that generative AI has historically had difficulty with.

Gemini Omni – a name that’s just as much of a mouthful as Nano Banana – can generate ten-second video clips, including audio, using existing photos, videos, text prompts, or a combination of those three. Google calls it “Like Nano Banana, but for video.”

The ability to generate videos from existing videos allows users to reimagine footage. Google says that Gemini Omni can transform a character, an object, or the environment “without ever losing the thread of your original scene." That even includes changing the camera angle, as well as the style of the video.

The launch notably comes after OpenAI shuttered its video generator Sora. Google says that Omni will replace its earlier video generator, Veo, inside the Gemini app.

One of the reasons that Google says that Omni is the video version of Nano Banana is that the new AI tool uses avatars. Like with Nano Banana, Google says that digital avatars allow creators to generate videos that still look like themselves.

I tried generating images of myself using Nano Banana when it first came out, and the experiment was both impressive and a bit terrifying. I still haven’t forgotten the images that resulted from that experiment.

The ability to generate videos that look like real people – and have a realistic look to them, as Google says – likely come with some risks. The launch is already raising questions, introducing more “AI slop” in an era where many social media users already complain about AI-generated content.

In a Reddit forum, one user said that making four videos took up an entire Pro Plan usage window. Another commented that “censorship is crazy tight,” suggesting there’s at least some guardrails in place to prevent AI misuse like generating fake political news or graphic content.

But, the slew of AI-related announcements coming out of Google I/O also includes more ways for users to detect if something is AI-generated. The Gemini app already has the ability to detect AI-generated images and videos that use the SynthID watermark, and Google is now rolling out the feature to Search. It will also be added to Chrome “in the coming weeks.”

The Omini-generated videos will come with a SynthID watermark, which means that a check in the Gemini App, Gemini inside Google Chrome, or a check using Google Search will indicate that the video is AI-generated.

You may also like

Ready to create real video? These are the best budget vlogging cameras.

I was fortunate enough to see The Rolling Stones in 2018 and I can tell you that they didn’t need ‘deepfaking’ then. It would seem that 82 is just a number for Mick, in particular, who is known for being incredibly dedicated to maintaining his fitness so he can strut his stuff when the band goes on the road. So, when The Stones released the music video for their latest track In The Stars, I was somewhat surprised to find that they’d been digitally de-aged.

As a bit of a music nerd, I’ve been trying to figure out exactly what period the music video is set. Ronnie Wood joined the band in 1975 to replace the outgoing Mick Taylor, so I think the mid ‘70s is a safe bet. So, let’s get the not-so-controversial bit out of the way first. In The Stars is an absolute banger. It really does sound like a forgotten deep cut from the band’s heyday.

I was invested right from the opening piano riff and falsetto chorus of “oohs”, but when the video cuts to Keith picking up his famous butterscotch Telecaster ‘Micawber’, and Mick pouts “One, two, free, and…” I knew I was in for a treat. The punchy riff that kicks in afterwards, guitars right on the edge of breakup, transports you right back to the ‘70s. And you know what else does? The youthfulness of Mick’s face.

In the video, Mick, Keith and Ronnie have all been transported back in time via the power of deepfake technology from Deep Voodoo, founded by none other than Trey Parker and Matt Stone, the creators of South Park. This is the same technology that was used on Kendrick Lamar’s 2022 The Heart Part 5 music video, transforming the rapper’s face into various likenesses including OJ Simpson, Kanye West, and Kobe Bryant.

It was also used to de-age Billy Joel in his music video for his first song release in over 15 years, 2024’s Turn the Lights Back On. I'm not an animation expert, I don't know how these deepfakes were created, but Deep Voodoo describes itself as an "artificial intelligence entertainment startup" and AI is heavily used in deepfake technology.

Anything AI related is going to cause heated discussions and I’ve repeatedly stated that I’m no fan of many generative-AI uses. As much as I like Kendrick Lamar, superimposing other likenesses onto another person’s face just doesn’t sit right with me. However, in the case of both Billy Joel and The Rolling Stones, I find the usage to be less offensive.

Retaining the human element

Firstly, the deepfakes are depicting the artists themselves. They’re fully aware of what’s being done to their likenesses. And that’s important, because a huge ethical question mark hanging over genAI is the scraping of other artists’ works. However, these deepfakes aren’t sampling creative works to form a rival creative work, they’re reproducing the likeness of a person who already exists. To me, that hits differently than, say for example, if the videos depicted a stylized, animated cartoon.

Cartoons have art styles, they require concept artists, animators, etc, so I’d question where that deepfake is getting its artstyle from. As an aside, I couldn’t help but notice that the likeness of The Stones’ late-great drummer, Charlie Watt, has not been reproduced for the music video. Now, I didn’t know Charlie personally and I don’t know The Stones, so this is pure conjecture, but I can only assume this was intended as a mark of respect.

Another big reason why I’m less annoyed by this usage of genAI is that all three videos have one thing in common. They are all real. They weren’t simply created using a prompt. Judging by the long list of contributors including DPs, directors, makeup artists, actors, etc, there is a strong human element to these videos.

The deepfakery was applied to real footage of real people and captured in a setup designed and produced by real people. I’ll leave you to watch the video and decide for yourself: A time-travelling celebration of The Stones, a creative faux pas or simply a step too far into the uncanny valley?

You might also like...

Thought the article was over? When you start me up, I'll never stop, so here are a few more article recommendations for you. On the subject of AI: Even Marvel artists aren't immune to being replaced by AI – and I believe Disney is on the cusp of another ‘Dark Age'. Why the AI ick means retro-style photos, videos and cameras are here to stay. Plus, an owl photo was dethroned from wildlife comp amid backlash: GenAI paranoia is ruining photography contests.

It was Star Wars Day on Monday and, as I sat watching the second season of Andor on Disney+, I felt numb. I don’t just like Star Wars, I love it. I’ve always loved it. But this was a May 4th I’d rather forget.

The previous evening, I’d watched an Instagram Reel released by Evangeline Lilly (embedded below) who played Wasp in the Marvel Cinematic Universe.

She commented on last month’s news that Disney had laid off 1,000 employees in a move that, according to Variety, CEO Josh D’Amaro said would “streamline” operations.

But Lilly, who had consulted friend and affected Disney employee, Andy Park, alleged that “the people who invented these characters in the first place, who designed them in the first place are now being replaced by AI.”

Now, let me be clear here. Disney has not attributed the layoffs to AI and in the video’s description, Lilly states: “***I can not confirm or deny that the motive for these layoffs was AI replacement.”

But it’s not exactly surprising that artificial intelligence has emerged as a keen subject of interest. This Guardian article highlights how Marvel Studios has come under the microscope for its treatment of creatives in the past.

And let’s not forget when Industrial Light & Magic revealed the horrendously tactless Star Wars Field Guide during a TED Talk this time last year.

A frankly embarrassing selection of clips of ‘alien creatures’ or what I consider to be an AI-generated mess of various animals spliced together to create a line-up of creatures that has more in common with a cabinet of curiosities than a galaxy far, far away…

You might be wondering what all this has to do with photography and videography. Well, everything. All creative industries are intrinsically linked and, should companies decide that artists are expendable, photographers and videographers (and journalists) should be very worried, too. I know I am.

Again, we don’t know exactly why Disney has downsized so rapidly – but I would like to remind Josh D’Amaro of a very troublesome period in Disney’s history.

Dubbed Disney’s Dark Age, the Seventies and Eighties weren’t kind to the House of Mouse. Various factors contributed to a decline in feature-length quality, most notably a loss of leadership due to Walt’s death, and the adoption of a faster and cheaper method of animation.

With an overreliance on live-action remakes and the directionless cash cows that both Marvel Studios and Star Wars have become, it’s hard not to feel like Disney is experiencing another creative slump. Getting rid of human creatives is not the answer.

This week my colleague, Alan Palazon, reported that Instagram is testing the ability for AI content accounts to label themselves as an “AI creator,” along with his opinion: “I think the platform needs to enforce it.” I couldn’t agree more.

And it’s not just the presence of unlabelled AI-generated content that’s a problem; the presence of genAI is affecting how people see photography, videography, art, CGI et al. Time and again I’m seeing photographs and other creative works being accused of being AI generated, when they’re genuine.

I’m not delusional. I know that AI is here to stay and I think it has the ability to bolster and embolden creatives. But genAI should never be considered as an alternative.

Huge creative companies like Disney pave the way – everyone else follows suit. I don’t think it’s hyperbole to say that the future of the West’s creative industry is – to a large degree – in Mickey Mouse’s gloved hands.

On the other hand, Jim Lee, CCO of DC Comics, spoke at New York Comic Con 2025, where he stated: “DC Comics will not support AI-generated storytelling or art. Not now, not ever, as long as Anne DePies and I are in charge.”

Disney, I have no interest in AI-generated slop. Let me know what you think in the comments below. Creatives, assemble!

You might also like...

Want to hear more of my ramblings on AI? I think restoring this iconic photo with AI is heresy and sets a dangerous precedent. I'm also certain that in this new world of AI, photography will live and die by how transparent we photographers choose to be about the images we create. And for something a little lighter: Nikon Coolpix S9300 compact camera is Gen Z approved. Bright, bubbly, cheap and… heck now I want one, too!

We all know that retro-style imagery and video footage are in right now, and I don’t see this supposed craze dying down anytime soon. If anything, I think people are going to be craving retro-style imagery increasingly and it’s largely because of AI. I hate to sound like a broken record, but AI-generated imagery is giving a lot of people the ick. In fact, only recently did I finish writing an article where I claimed that GenAI paranoia is ruining photography contests.

In a world where social media is perpetually drenched in a soupy broth of AI slop, retro-style imagery feels safe, genuine, real. But genAI isn’t to blame entirely. Before the generative boom, we had already reached image-fidelity oversaturation.

The megapixel race had become a bore, the very latest editing trends dominated Instagram, and computational camera phones were presiding over an era of headache-inducing sharpness and HDR imagery that made even family snaps look like a commercial advertising campaign. In short, the industry was ripe for a retro style reckoning.

Cue influencers flocking to buy up retro compact cameras and inflating prices, Fujifilm’s penchant for retro cameras bolstering its brand to the forefront of Gen Z’s radar, and a renewed interest in film photography. Heck, last week, I wrote about the Nikon Coolpix S9300, a camera I forgot existed the day after its release way back in 2012, yet it was trending!

And let’s not forget society’s lust for retro style in general. Throwback entertainment such as Stranger Things and Guardians of the Galaxy, not to mention retrospective fashion trends. My hope is that we’ll continue to search for ways to stay more connected to image making, beyond our computationally driven camera phones and soulless genAI prompt text boxes long into the future.

If I’m right, vintage cameras and retro-style cameras will continue to sell as a bastion of authentic image making. But if I’m wrong, and retro style is largely superficial, I guess it won’t be long until people start turning to ‘low-fi’ AI imagery. The technology already exists.

Only time will tell.

You might also like...

Into authentic retro style photography? Look no further than the best film cameras. If you like the retro look but don’t want to give up mirrorless mod cons, here are the best retro cameras. And if you just want the most advanced darn tech on the market, here are the best mirrorless cameras.

I’ve had the pleasure of judging a lot of photo competitions in my time, and I can tell you that the most contentious category to critique by far is wildlife. That’s because it’s fraught with potential rules violations that are often extremely difficult to investigate: captive subjects, baiting, harassment, compositing, and more recently, generative AI.

So you can imagine the furore when the National Wildlife Federation selected a stunning image of a great horned owl, perched in front of an aurora in Oklahoma as the winner of its Garden for Wildlife Photo Contest. It didn’t take long for the comments to whip up a storm. The original Instagram post has since been removed, but PetaPixel reported that these included accusations of both AI involvement and compositing.

Particular attention was directed towards the owl’s talons, which appear to show too many digits – often a hallmark of generated AI imagery. And while the aurora has been explained away as a geomagnetic storm, a well-lit image of a wildlife subject in front of a well-exposed night sky would be extremely challenging to achieve in a single frame.

This is an important point, because the Garden for Wildlife Photo Contest does not allow composite imagery. And indeed, to the competition's credit, the image has since been disqualified due to it being deemed a composite and an incredible image of a Yellow garden spider by Nicole Land has been announced as the deserved winner. As far as I can see, to date, the competition has offered no conclusion as to whether or not generative AI was involved in the offending image's creation.

Now, I’m generally very sympathetic to competition judges. Frankly, it’s a thankless job and you can’t please everyone. Heck, it’s pretty hard to please yourself, because when you’ve sifted through thousands of images and ended up with tens of potential winners, you rarely feel wholly satisfied with your choices.

And situations like this aren’t entirely on the competition. Images are judged in good faith, that’s why there tends to be a very carefully worded list of rules. But in this instance, I cannot understand why the offending image wasn’t flagged, immediately. To me at least, it screams further investigation.

However, this entire situation has highlighted another ugly side of genAI imagery. And I don’t believe it’s being talked about enough. AI accusations. I don’t know whether or not the image in question was partially or entirely AI generated or not. But it still breached competition rules and in doing so it still caused an AI uproar.

Every time something like this happens, people get a little bit more paranoid about AI, to the point that we now live in a world where photographs, videos and artworks are constantly assessed for even the slightest hallmark of genAI involvement.

Whether or not the content in question is genAI is kind of superfluous to my point. The rot has already set in. The world is more skeptical of photographs and videos than ever before. Heck, you can’t even come across a funny video of a cat – mid doomscroll – without ditching the laughter for a second to really analyze the footage.

Hmm, is that AI? Or is that…

Too late, the moment – the impact – it's gone...

You might also like...

Interested in wildlife? Here are the best cameras for wildlife photography and the best lenses for bird photography. Plus, check out my 10 tips for wildlife photography.

VSCO has carried out a study on AI usage in photography – and the results suggest that a large majority of photographers are using AI at some point in their workflows.

VSCO surveyed 401 photographers, both professionals and hobbyists, on how AI is perceived and applied across creative and business workflows, and found that 83% of photographers in general use the technology.

What’s more, of the professional and hobbyist respondents, 68% and 34% respectively said they use AI on at least a weekly basis. VSCO says the growth in AI adoption spans multiple genres from wedding photography to landscapes.

The survey doesn't mean that 83 percent of photographers are using generative AI – nearly half of the respondents reported using AI for repetitive, mundane tasks. Around 63 percent are using general tools like ChatGPT and Claude. Assisting with editing and culling is the most popular role for AI, followed by shot planning, business administration, and mentor-like feedback.

Honestly, I’m not surprised to see such a high percentage of photographers using AI, and I think it’s only going to grow when you consider how much time it can save you in post-production.

In fact, the study also revealed that 38% of photographers used AI more in 2025 than the year before, and 29% started using it for the first time in 2025, while just 2% of photographers have “pulled back” on AI in the past year. However, VSCO didn't specify to what extent “pulled back” meant.

Opposition to AI in photography is mainly geared toward enabling it to take the creative reins. However, it appears that most photographers use it to handle the more repetitive and less-creative aspects of the craft, as the study showed 55% of photographers use AI to assist with editing, culling, and to automate aspects of post-production.

I’d say the study highlights, at least to some extent, how photographers are increasingly accepting the benefits of AI rather than seeing it as a threat to their jobs. Respondents were also asked how they feel about AI, and 49% of professionals and 37% of hobbyists said they were curious about using it.

Of course, there are still what I’d call severe concerns surrounding AI and its adoption within photography, with 17% of the survey respondents indicating they felt skeptical and 5% saying they felt threatened by it.

Looking ahead, I can’t see future studies contradicting this one by VSCO, especially when I think about how I use AI to streamline my own post-production workflow and the hassle it saves me.

While this study certainly doesn't represent the entire photography community, I think it further highlights the inevitability that AI will become a staple of the craft, with the extent of this depending on individual photographers.

You might also like…

I'm sorry to break this to you, but your mirrorless camera probably already has AI built into it.

Generating video with AI typically involves typing in a prompt – and then starting over if the AI isn’t quite right. But software giant Adobe is building an AI that can adjust videos as they are being generated, allowing the user to change the camera angle or move objects around in the scene.

Adobe MotionStream is an experimental AI video creator that gives users real-time control over elements in the scene. The experimental technology allows users to change the camera angle and direct the movement in the scene using drag-and-drop style tools, along with sliders.

MotionStream, which was unveiled on April 10, generates the video in pieces in order to allow users to interact with the video near real-time.

All of this happens, Adobe says, as the video is being generated. MotionStream begins with a text prompt, but then offers “immediate visual feedback,” Adobe says the user can then move around the camera for a different angle, or move objects in the scene, watching the results adjust as those changes happen.

Adobe says that MotionStream is also designed to replicate more natural physical movements, which has historically been a challenge for AI.

“That’s where a lot of the magic happens—in the secondary effects that are really hard to control manually,” said Adobe Senior Principal Scientist and MotionStream researcher Eli Shechtman. “If you want to move an elephant, for example, you can click and move its body, but it’s a lot of work to manually make those movements look natural.”

“This currently requires skills and specialized software to rig, and animate or keyframe the animation, following a process that typically takes hours, if not days, depending on scope,” Shechtman added. “Instead, the underlying video generator behind MotionStream is basically simulating the world in real time. So, the elephant’s legs move naturally, and the ears flap naturally as the elephant moves. The model provides you with knowledge about the world, and you can interact with it.”

Generating videos with AI typically involves a wait, but MotionStream actually started from Adobe Research efforts to speed up the process by breaking it down into pieces. Users see the first piece while the latter pieces are being generated in the background.

“The natural next step, once we started breaking videos into pieces,” says Adobe Senior Research Scientist and MotionStream collaborator Richard Zhang, “was to ask for feedback from users as the video is being generated. That’s what brought us to MotionStream. It’s the fruit of a long line of research.”

Once the researchers overcame the delay, they worked to give users the tools to make adjustments, allowing users to interact using tools as the video is being generated to fine-tune the results.

MotionStream is, for now, only an experimental program, but Adobe is now sharing a preview of the program to the public.

You may also like

Browse the best video editing apps.

Humans have orbited the moon for the first time in more than 50 years. But besides the advances in space tech between Apollo 11 in 1969 and Artemis II in 2026, there’s another key technological change between those two launches: Artificial intelligence.

According to fact checks from mutlipe news outlets, including BBC and CBC, viral photos and videos circulating online claiming that Artemis II footage is fake are themselves AI-generated fakes.

Photos from TikTok show the four astronauts wearing a harness system in front of a green screen. But when the BBC ran the photos through Google’s SynthID AI check, which reads embedded watermarks on Gemini’s AI creations, the tool said that those green screen photos were generated with Google AI.

Eagle-eyed viewers watching a CNN interview with the astronauts that aired earlier this week found unusual text overlaid on a floating gravity toy named Rise, leading some to call the video out as an AI fake. However, an analysis showed that those artifacts didn’t appear on the original footage – BBC’s verify team says that the unusual text artifacts were not a green screen error or a sign of AI, but a glitch in the tool used to place text over video footage recorded in camera.

One survey of US adults from 2021 suggests that 12 percent of Americans believe that NASA did not land on the moon in 1969, while another 17 percent responded as “unsure.”

Fake photos being used to claim a historic event was faked is ironic, yes, but also a key example of living in the AI era. Photographs could be faked before the age of AI, but the ease at which anyone with a keyboard can create a fake image is increasing distrust over the credibility of photographs. One survey from 2025 found that 82 percent of respondents said their confidence in media has decreased as a result of AI-generated content.

NASA has been sharing photographs taken on the Artemis II mission since the Orion launched on April 1. Metadata from photographs shared from official NASA channels show photographs that were taken with the Nikon D5 DSLR, the Z9 mirrorless camera, an iPhone 17 Pro Max, and even a nearly 12-year-old GoPro.

You may also like

Browse the best mirrorless cameras or the best DSLRs.

As a freelance creator, I’m profoundly opposed to Generative AI. But when I object to people using it, I’ve been accused of being ‘elitist’. “Not everyone is a great photographer, or a great writer”, they argue. “Why shouldn’t people use AI to help them do things they can’t already do?”

While I admit some AI-powered tools can be useful – such as using Lightroom masks to enhance an image, or a grammar app to correct your writing mistakes – I’ve noticed that the power of Generative AI to do everything has led to a curious air of entitlement in some quarters.

I’m not talking about individuals with disabilities, for whom AI tools can help level the playing field. This is about people who want to cheat their way to becoming ‘experts’ overnight, without putting in any of the hard graft.

One online commentator summed it up when they said using Gen-AI is “the best way to sound like a professional writer without being one.” Never mind that to become a professional writer takes skill, dedication, and practice.

It’s the same with photography. While professional photographers have spent years investing in and honing their craft, so-called ‘AI artists’ now feel entitled to leapfrog their way to the top, using resources trained on images stolen from the very experts they’re now trying to replace.

But if you can’t do something because you don’t have the skills or you never bothered to learn, why should you suddenly get to reap the rewards? I can’t do plumbing, so if I have a leaky tap, I either need to learn how to fix it or hire an experienced plumber. If you need a photo and can’t take one yourself, hire a photographer, or license a stock image (which isn’t even expensive).

Otherwise, it’s just theft and cheating; a way to bypass the experts, and fake skill and experience without actually having any.

It seems the old expression of ‘Those who can, do; those who can’t, teach’ has been updated for the modern era. Those who can still produce quality, authentic work. Those who can’t, use Generative AI.’

Aftershoot has pledged that its software will remain an “assistant” to photographers rather than a “replacement,” responding to what it has described as growing fears that AI could eventually overtake the industry.

In a newly published webpage, the company behind the AI-assisted image culling tool of the same name has emphasized that “the camera will always need you” and reinforced this stance with a three-point commitment to photographers.

Aftershoot has promised that its software will never introduce features that take over creative control, that users will always have the option to opt out of sharing data for AI training, and that decisions about the platform will remain community-driven.

This announcement comes just a few months after backlash against Evoto — the company behind AI-powered desktop editing tools Evoto AI and Evoto Video — which recently piloted an AI headshot generator capable of creating images without the need for a photographer.

In an accompanying press release, Aftershoot said the biggest concern within the photography industry today is AI replacing photographers. It added that while many AI companies developing tools for photographers have largely avoided addressing these fears while benefiting from the AI boom, Aftershoot is choosing to “break the silence.”

Co-founder Justin Benson addressed this directly in the press release: “I got into photography because I love it. Then I joined to build Aftershoot because post-processing was eating the time I had to actually shoot. And that's still why we're here. Not to replace what you do.”

According to Aftershoot, nearly 250,000 photographers use its software to streamline post-processing. The company’s commitment to protecting photographers’ jobs will surely strengthen its relationship with users and may set a precedent for how developers of AI tools work with the photography community going forward.

You might also like…

Read the full Aftershoot review. Or, read about the best photo culling apps.

In an unexpected move, OpenAi has announced that it’s pulling the plug on Sora, the company’s AI video generator.

In a post on X, OpenAI announced that it is shutting down Sora, a video generator that has now been around for more than a year. “We’re saying goodbye to Sora,” the post reads. “To everyone who created with Sora, shared it, and built community around it: thank you. What you made with Sora mattered, and we know this news is disappointing.”

The AI giant did not offer a reason for the shutdown. But, earlier this year, ChatGPT started including ads on its chatbot amid pressure to generate revenue. According to The New York Times, OpenAI is spending “tens of billions of dollars” to build the infrastructure needed for AI expansion. Just a few months ago, OpenAI and Disney announced a $1 billion partnership.

An AI video generator is typically more resource-intensive than a chatbot, so the costs of running Sora were likely very high. And, Sora was a free video generator platform.

While the closure of Sora seems unexpected, the news may be a sign of hope for videographers creating videos with cameras rather than computers, as Hollywood has expressed concern that Sora posed a "significant risk" to creators. Many creators consider generative AI a threat. If one of the largest AI software companies can’t turn a profit from a video generator, some creators may take that as a positive sign.

The news comes amid rising costs of storage due to the data centers needed to run AI platforms. Some experts have hinted at an AI overinvestment, while others warn of an AI slowdown or even an AI bubble burst.

Of course, Sora isn’t the only video generator, but the shutdown leaves one less free video generator option. Other AI platforms capable of video generation include Adobe Firefly and Google Gemini’s Veo.

You may also like

Take a deep dive into the best cameras for video, or the best video editing apps.

Adobe has launched a big Firefly update in a bid to make the all-in-one creative AI studio more personal to you. If you’ve ever used Firefly, or any AI image generator for that matter, you’ll know that while a prompt allows you a great deal of control over the subject matter, it’s a lot harder to convey the style that you have in your mind’s eye. That’s where Firefly’s new Custom Models feature comes in.

This function, now in public beta, allows users to train the artificial intelligence to replicate their own personal art style when generating images and is optimized for illustration, character, and photographic styles. All you have to do is feed the AI 10 to 30 of your own images to create a custom model. However, the quality of the training data matters, so Firefly provides you with a ‘Model Score’, telling you how effective your selection of images is at training your custom model.

Adobe tells me that the custom models are built on Firefly Image Model 4, with plans to implement Image Model 5 in the future. Not only will the model help you generate consistent-looking imagery, but it will also automatically generate tags and captions for your images. I got to sit in on a demo by Adobe’s Principal Director Evangelist, Paul Trani, who demonstrated his own custom model and explained how he uses it primarily for ideation, as well as sourcing components for his composite artworks.

No doubt the questions on most creatives’ minds will centre around privacy and copyright issues, but Custom Models is designed for commercial use. Adobe’s Vice President of Product Marketing for Creative Professionals, Deepa Subramaniam, made it very clear to me that “Your data is yours.” That means user data is not used to train Adobe’s models, and Adobe doesn’t claim ownership over user content from Custom Models, Firefly models, or third-party models.

However, Custom Models does raise the issue of users feeding Firefly content that they don’t own. Thankfully, Adobe is intent on combating this. According to Deepa, “When you’re using custom models and training custom models, it’s very clear that we expect you to have ownership over the content that you’re uploading and that it be yours.”

And indeed, Adobe regularly reviews users’ custom models and will remove those that don’t play by the rules. Content created by a custom model will also have that model embedded in the content’s metadata. Deepa also added that “this is definitely an area that we’re using the public beta to also get feedback on, ways to improve those protective measures as we look towards GA.”

As it stands, Custom Models is rolling out via Firefly Boards and Firefly Generative Image. However, Adobe is keen to point out that creatives are using more generative credits than ever and that, according to its recent survey, 86% of creatives use generative AI. And while Firefly is the home of Custom Models for now, there are plans to implement the technology within other Creative Cloud applications in the future.

It’s not all about Custom Models, though. Adobe has also introduced a Quick Cut video feature, which is said to organize video footage into a “structured first cut in minutes”. Adobe is also expanding access to its Project Moonlight private beta, which centers around agentic AI assistants that are designed to make prompts less passive and more active.

Instead of writing a prompt and setting the genAI to work, Project Moonlight’s AI assistants are designed to work with creators to deliver their vision, turning prompt generation into a conversation. According to Adobe: “Moonlight gets you. It understands your style and gives you control over the work, drawing from your own assets and libraries. You become the creative director of your own world”.

You might also like...

Check out the best photo editing software and the best photo organizing software. Want the latest Adobe news? Adobe tells creators "At the end of the day AI is just a tool" – BrandonB and Niko Amolina on stage as Adobe holds first-ever Creator Live event in London.

Sony’s autofocus system may actually hold the tech to prove whether or not a video originated in camera – or is an AI fake. This week, the tech giant expanded its Camera Verify AI detector to support videos as well as stills, but the way the software works actually uses some 3D tech likely tied to Sony’s autofocusing system.

Camera Verify is a beta software that analyzes photos – and now video – to determine whether or not that content was created inside one of Sony’s cameras in order to flag AI fakes. The program is designed for journalists and news organizations and requires a paid license for the software. The program launched last year for still photos, but now supports video.

The program uses three key verification factors, including reading digital signatures under the C2PA (Coalition for Content Provenance and Authenticity) and verification using the capture time stamps.

But the third factor that the software uses to prove a video came out of a camera and not a bot is actually tied to the brand’s autofocus system. Sony explains that the software also looks to see whether a real-world subject was captured by using 3D depth detection.

Several of Sony’s high-end cameras have a feature called Focus Map. It’s designed to help videographers quickly see what’s in focus and what’s not. But the depth data has more than one purpose, and Sony is now using it to help fight fakes by looking for the depth data inside the file’s metadata.

Without a 3D depth map, nefarious users could potentially use a camera to photograph an AI-generated image or video playing on a TV screen in order to get that C2PA camera stamp. But by using a 3D depth map, Sony’s software can also tell if the image or video is a shot of a 2D print or screen, or whether it’s actually a 3D scene that happened in real life.

While Camera Verify is only in beta and only available for certain journalism outlets, it’s interesting to see autofocus tech being used to fight AI fakes. Sony's autofocus uses AI for subject detection, so it's a bit of AI fighting generative AI. The new video detection capabilities are rolling out with support for footage shot on the Sony A1 II, A1, A9 III, A7R V, A7S III, and A7 IV. Support for the A7 V is also expected to be added as early as May 2026.

You may also like

Browse the best Sony cameras or the best Sony E-Mount lenses.

The above photo was taken in July 1917 by a teenage girl, Elsie Wright, in the garden of her Cottingley home in Yorkshire. It shows her younger cousin, Frances Griffiths, surrounded by dancing fairies. The fairies were paper cut-outs, secured to the ground with hat pins. And yet the image fooled millions of people including, most famously, Sherlock Holmes creator Arthur Conan Doyle.

As photographers, this story should make us feel simultaneously vindicated and uneasy. Vindicated, because we understand better than most how profoundly a camera can deceive. Uneasy, because the lessons this affair teaches have never actually been learned; and the same failure is playing out right now, at unprecedented scale, with AI-generated deepfakes.

Last week, an episode of BBC Radio 4's The Long View drew the parallel with characteristic sharpness. As it pointed out, the number of deepfakes shared online rose from around half a million in 2023 to eight million in 2025. The Cottingley story, it turns out, is not ancient history. It's more of a manual.

What Kodak missed

Here's the detail that should interest photographers most. When Doyle set out to verify the Cottingley fairy photographs, he did not rely on instinct. He sent the images to Kodak, who confirmed there had been no double exposure. He also commissioned Harold Snelling, an independent expert in photographic manipulation, who reached the same conclusion.

Both experts were looking for darkroom trickery. Neither considered the possibility that the deception had happened entirely in front of the lens: physical cut-outs, carefully positioned, photographed straight. The technique required no manipulation of the negative whatsoever.

It is a forensic lesson that echoes loudly today. When we try to detect AI-generated images, we look for the artefacts we know about: the wrong number of fingers, the inconsistent lighting, the smeared background text, the tell-tale compression signatures of GAN-generated faces. But the deception, as in 1917, tends to stay one step ahead of the verification method.

New tech, old story

What the programme makes compellingly clear is that this is not really a story about photography or AI. It's a story about a very old human habit: we grant uncritical trust to whatever our newest representational technology produces, and we keep doing it until something goes badly wrong.

In 1917, photography was still close enough to its origins to feel miraculous. The cultural idea that the camera could not lie was close to absolute. At the same time, spiritualism was flourishing in the grief-saturated aftermath of the First World War. Conan Doyle had lost his son. He desperately wanted these photographs to be real, even if he didn't consciously realise it.

It gets worse: Conan Doyle never knew that Snelling made adjustments to the photographs (improving exposure, adding detail to the fairies' wings) before they were published. The image he championed in The Strand Magazine in 1920 was not the image the girls had taken. The chain of custody was broken before he ever saw a print.

As BBC disinformation specialist Marianna Spring observes in the programme, the mechanics are similar today. Deep fakes succeed not because they are technically flawless – they're often very much not – but because they confirm what the viewer already wants to believe. 

The lesson that never lands

The woman who'd taken the Cottingley fairy photos finally confessed to a journalist in 1983, more than 60 years after on. Her explanation for the long silence was not shame but consideration: she and Elsie did not want to humiliate Conan Doyle and his associates while they were still alive. The confession came too late to do much good, though. By then, the story had become myth.

So what can photographers—people who spend their working lives understanding exactly what a camera can and cannot show—usefully take from all this into the deepfake debate? I'd say the lesson is this. 

The Cottingley photographs were verified by photographic experts and still fooled the world, because the experts were solving the wrong problem and the audience had already decided what it wanted to see. Technical literacy matters, but it will only get you so far when motivated reasoning is doing the heavy lifting on the other side.

The most important lesson from 1917, then, is also the most uncomfortable one. Before you try to spot whether an image is fake, ask yourself whether you, or others, want it to be real. That question will tell you more than any techno fix or detection tool currently on the market.

A few days ago, the Instagram algorithm served me a gorgeous image of a leopard by a photographer I don’t follow. I clicked the caption to find out more. But instead of an explanation of where or how the photo was taken, I was slapped in the face by a tepid stream of AI-generated waffle.

You probably know the style. Quasi-poetic word vomit like: “the animal moved effortlessly through a canvas painted by light and pure chance.” Or this, which accompanied another picture of two penguins: “their paths crossing and uncrossing, a quiet choreography of pursuit and promise.”

I understand why, if you’re not a confident writer, you might be tempted to get ChatGPT to write your captions for you. But if you do, you’re shooting yourself in both feet at point-blank range for several crucial reasons.

First off, generative AI is the mortal enemy of creators. Where once a company would hire a photographer, a copywriter and a designer to make an advertisement, now one person can chuck in a prompt and it’ll spit out something usable – stolen and regurgitated from the very people whose jobs it is replacing.

Why would you want to feed the beast that’s eating you alive?

Next, AI is known to have a colossal environmental impact through building data centers, mining of raw materials and sucking up scarce water resources. For wildlife and landscape photographers to embrace it, without a second thought for the nature they claim to love, is a baffling contradiction.

But most of all, AI-generated captions aren’t even good. They’re bland, meaningless and boring: the written equivalent of chewing on polystyrene packaging. They’re also completely unnecessary. I’m not sure who told photographers that they need to become poets as well, but trust me – they really don’t.

If you’re a photographer on Instagram, your followers want to see your photography. They don’t expect you to be a great writer. They don’t want a novel, they just want to know where you were, how you took the image and what you thought. That’s all.

Don’t believe me? Think I’m a Luddite, the modern-day equivalent of a typist objecting to the invention of the computer? It seems I’m not alone. When I posted about this on Instagram, I received nearly 1000 likes and over 100 comments agreeing that AI-generated writing is “obvious, boring and harmful.”

In an era where trust and authenticity are paramount, your genuine voice, even if it’s not perfect, is worth so much more than a bowl of bland, robotic porridge.

After all, if you’ve clearly used AI to write your caption, it makes me wonder if you used it in your images too!

Bella will be talking on the Canon Spotlight stage at The Photography & Video Show on Tuesday March 17 at 1pm

Save 20% on your Photography & Video Show tickets with code DCW26

You might also like…

Check out the best cameras for wildlife photography and the best lenses for bird photography and wildlife.

Since AI has become commonplace in smartphone cameras and even the dedicated camera called Caira, photographers have been clutching our cameras that spit out un-generated images a little tighter. But here’s the thing: most mirrorless cameras already have AI inside.

I was reading an opinion earlier this week that predicted that AI would come to mainstream cameras, but that such a move would come with dramatic pushback by photographers. I agree to an extent…except that AI is actually already inside most mirrorless cameras.

See, there’s an important distinction in the AI era: AI and Generative AI are not on the same ethical playing field. AI simply means software that “learns” from a set of training data. Generative AI takes that ability to learn and “creates” something from that data: a photo, a video, a song, an essay.

Cameras already have AI built into them – and they have for years. The biggest culprit? Eye detection autofocus and subject detection autofocus are AI technologies. Canon’s in-camera upscaling and noise reduction is also AI; it’s just not generative AI.

But, using AI software to focus on the eye of a bird in flight isn’t the same thing as using AI to generate an image. A generated image doesn’t depict reality, whether or not it was generated on a computer or partially inside a camera.

I don’t think it’s a coincidence that compact cameras are returning as smartphone cameras become increasingly reliant on computational photography and generative AI. I want to freeze memories, not generate them. And if my memories aren’t perfect, why should my photographs be? Give me a compact camera over a smartphone’s overly perfected images any day.

I have no qualms with an AI that can help me lock the focus in quickly on a camera that still freezes real memories. Such a system is built on real images, yes, but the tech doesn’t compete with or attempt to replace photographers – just makes the job a little easier.

But a camera that not only makes things up but does so as AI companies make billions of dollars on stolen images without paying the original creator? That I have a problem with, and I know I’m not alone.

The question that remains is this: will major brands eventually integrate generative AI into cameras? It's certainly possible. Canon and Panasonic recently invested in a generative AI startup that generates lifestyle product photos. A start-up has created Caira, a mirrorless camera that attaches to a smartphone and has Nano Banana built in.

But integrating generative AI into a camera from a major brand would alienate the very people who buy cameras: photographers.

You may also like

Annoyed at modern tech? These are the best DSLRs that you can still buy.

Adobe Firefly, the software giant’s AI platform, can generate videos – but now the platform can also create an edited video from real camera footage. Firefly Quick Cut is an AI-driven tool that gives creators a jump start by creating a quick draft edit.

The new tool works by asking videographers to upload their clips and then describe what they want to create in a text prompt. Adobe notes that the instructions can be as simple as what type of video you want to create, such as a product review or interview, or as precise as uploading an entire script to work from.

Quick Cut comes with a few built-in tools to help guide that initial edit. Creators can choose the aspect ratio and set the pacing, or even set a specific final length for the video. Adobe notes that creators can also upload B-roll to a separate track in order to clearly identifythe main footage from the supporting shots.

Quick Cut is designed to be more of a starting point than a replacement for traditional video editing. Adobe describes it this way: “We designed Quick Cut to give Video creators a clear starting point they can shape, refine, and make their own. It’s a fast way to get from ‘I have clips’ to ‘I have an edit I can work with.’”

After the initial Quick Cut, videographers can then refine the results.

The tool is rolling out to Adobe Firefly, the software giant’s online-based AI platform, beginning today, February 25.

Adobe Firefly is free to try, but the free version limits the features and number of generations. Adobe is currently offering unlimited image and video generations up to 2K for Firefly Pro, Firefly Premium, and customers using the 4,000, 7,000, or 50,000 credit plans until March 16.

You may also like

Browse the best video editing software.

Video streaming giant YouTube has begun to crack down on channels potentially earning millions of dollars publishing “AI slop” content created solely using artificial intelligence (AI).

In a recent move, the platform either completely removed or wiped content from 16 leading AI-driven channels which, collectively, had amassed over 35 million subscribers and 4.7 billion views.

In an open letter to the public published in January this year, YouTube CEO, Neal Mohan, said the platform is “actively building” on established systems that have been “very successful” in combatting the spread of “low quality AI content”.

The action comes after a report published by the video creation platform Kapwing identified the “AI slop” YouTube channels with the most subscribers and their earning potentials, thought to be in the millions.

“AI slop” is widely regarded as content generated entirely using AI tools that's low-quality and repetitive. AI slop channels typically publish new posts at much higher rates than human-content-led channels to maximize engagement and advertising revenue.

Also called ‘AI brainrot”, videos often include historical themes littered with inaccuracies and surreal, nonsensical scenarios, typically reusing the same script, music and visuals.

One of the leading AI slop channels identified in Kapwing’s report which subsequently had the majority of its content wiped was Three Minute Wisdom (1.7 million subscribers / 2 billion views).

According to the report, the South Korean channel, which initially had 140 videos depicting “photorealistic(ish) footage of wild animals being defeated by cute pets”, but now only shows 20 live videos, was potentially earning $4,036,500 in annual ad income.

Other AI slop channels that have been completely removed from YouTube following Kapwing’s report include CuentosFacianantes (FascinatingTales), which published “low-quality Dragon Ball-themed videos", and Imperio de Jesus (Empire of Jesus), a religious channel which had over 5.8 million subscribers.

However, not all of the lucrative AI slop channels mentioned in the report seem to have been penalized by YouTube.

Bandar Apna Dost, an Indian channel with over 3 million subscribers and 2 billion views, and an estimated annual ad revenue of $4,251,500, still has over 400 videos live.

Kapwings report also calculated that over 20% of videos shown to new users on YouTube are AI slop. But as yet, it’s uncertain how YouTube plans to strike a balance between promoting AI-generated and human-crafted content to viewers.

YouTube CEO Neal Mohan previously stated that generative AI has been “a revelation”, saying that it can do for video what the synthesizer did for music.

You might also like

Could generative AI hollow out the photo industry within five years?

Generative AI is often viewed as a threat to traditional photography, but two venture funds with ties to camera companies have recently invested in a startup that generates images from initial product photos. Canon and Panasonic have both invested funds in AI Model Inc., a start-up that generates images from basic product photos.

AI Model Inc. is a company that’s building technology for apparel companies to generate multiple types of content based on a basic product photo. The AI uses those initial images to generate images and videos of the apparel being worn.

Canon Marketing Japan MIRAI Fund, a fund managed by Global Brain Corporation in partnership with Canon Marketing Japan (a subsidiary of Canon Inc.), has invested in the latest round of funding and has also previously supported the start-up in 2024.

Panasonic Global also announced that the venture fund managed by Panasonic and SBI Investment Co, the Panasonic Kurashi Visionary Fund, has made an investment in AI Model Inc.

The investment amounts were not disclosed.

In one statement, AI Model Representative Director and CEO Daiki Taniguchi describes the start-up’s approach this way: “AI model is a platform that enables companies to generate and manage customized AI models and talent for their businesses, allowing them to rapidly and flexibly optimize creative content across a wide range of channels—including TV commercials, infomercials, banner advertisements, and in-store POP displays—without the need for photography, while maintaining brand identity and compliance.”

But, when Digital Camera World reached out to the respective companies, asking why a camera company would invest in generative AI, AI Model Inc. clarified how photography is part of the process. "AI models are not intended to completely replace photography," a representative said. "Instead, they are used as a complementary workflow to improve the efficiency of shoot preparation, enable the generation of additional images, and support low-cost."

"The process still requires capturing products worn by staff or mannequins using professional cameras, making high-quality photography an essential foundation," the AI company added. "In fact, photographers can focus more on product-focused shooting, where core skills such as lighting, composition, and accurate texture representation become even more important. By integrating photography with AI technology, both the quality and efficiency of creative work are enhanced."

In 2025, roughly half of global venture funding went to AI start-ups, according to Crunchbase. Investments in AI software companies have been the most common type of corporate investment for three years running.

“By combining AI Model Inc.’s AI-driven content creation technology with the imaging technologies and expertise cultivated by the Canon MJ Group, we are confident that we can create new services and contribute to the advancement of society. Based on this conviction, we decided to make the investment," Canon Marketing Japan told Digital Camera World.

“As AI is adopted across all industries, we believe that the development and effective use of advanced AI will significantly contribute to providing customers with new value and expanding future growth opportunities for our business,” said Kunio Gohara, the General Manager for the Corporate Venture Capital Office at Panasonic. “By combining AI Model's proprietary generative AI technology with Panasonic's deep insights into people's lifestyles cultivated over many years, we hope to create a future where we can communicate lifestyles tailored to each customer, based on how they envision a rich lifestyle.”

In a press release, AI Model Inc. says that AI will not take away jobs. “Our goal is not for AI to simply replace the work of creators, models, and talents, but to create new opportunities through co-creation and broaden the scope of creativity,” the press release reads. The startup also notes that it is working on creating a foundation “that allows companies to use AI models with peace of mind by formulating guidelines and ensuring transparency.”

You may also like

Browse the best cameras for product photography or the best professional camera.

Five organizations are lobbying in London this week, urging government action against what the group calls “the greatest acts of theft in modern history:” Generative AI.

This week, the UK’s Association of Photographers (AOP), alongside the Society of Musicians, the Society of Authors, Equity, and the Association of Illustrators, have joined together in a report urging legislative action to protect creative industries in the era of AI. The report, called Brave New World? Justice for Creators in the Age of Generative AI, comes as the organizations lobby in London to protect the UK’s creative industries.

The report combines data from across several creative industries, including the AOP’s recent survey on copyright and AI, with surveys representing 10,000 creatives from organizations that support upwards of 80,000 people.

According to the AOP, around 58 percent of photographers have lost work to generative AI. The reported financial losses are a 142 percent increase from losses reported in the January 2025 survey, amounting to an average of £35,000 (around $48k) in annual losses.

Comparing the AOP survey data to the organization’s surveys from previous years highlights how rapidly GenAI has affected the photography industry. In September 2024, 30 percent of photographers reported losing assignments to GenAI. By February 2025, that number had risen to 58 percent.

The AOP survey data also suggests that many photographers are reducing the number of photographs shared online, with the number of shared images dropping by 46 percent compared to last year. The Brave New World report suggests that more creators are keeping work offline to avoid being picked up and used to train generative AI.

The issue isn’t just a creative one but an economic one, the report suggests. In Brave New World, the AOP warns that without regulation, GenAI could “hollow out the UK’s £2.4 billion photography industry within five years.”

Across all the creative industries represented in the report – which includes photographers as well as musicians, illustrators, and writers – the report highlights a disparity between UK jobs in creative industries and those in the AI tech sector.

The report estimates that the creative sector is worth about £124.6 billion annually and more than 2.4 million jobs, which is almost 30x larger than the UK’s AI sector, with 86,000 jobs and £11.8 billion in 2024.

Creatives are also concerned that GenAI will reduce the number of entry-level and mid-level positions for new artists who are just embarking on a creative career.

In the AOP survey, 94 percent of photographers said they want to see a “Personality Right” introduced. A Personality Right refers to an artist’s ability to protect their name, likeness, and “distinctive creative identity.” In one of the report’s case studies, Tim Flach, President of the Association of Photographers, found that using the prompt “Tim Flach Tiger Image” generated an image similar to his.

The organizations are calling for a new framework for AI legislation modeled after the acronym CLEAR that, at a minimum, requires:

• Consent first: Creators across industries are calling for consent for training, including the ability to choose which works are used in AI training and which are not.
• Licensing, not scraping: The report calls for licensing to be developed in collaboration with relevant trade unions, rather than a one-size-fits-all model across all creative industries.
• Ethical use of training data: The report calls for creative works to be recognized as such and not treated as “free raw materials.”
• Accountability and transparency: The organizations call for standards that require GenAI to disclose what data was used for training, offer tools for creators to check to see if and how their work has been used, and to publish information about training data sources and licensing status, among other standards.
• Remuneration and rights: The standards call for payment for works used in AI training.

The report stresses that “creators are not anti-AI. They want to participate in its economy, not be erased by it.” Licensing data sets for AI is an industry valued at $2-3 billion a year and expected to grow, the study suggests.

The CLEAR framework is both a call for government action as well as a guideline for creative industries and organizations to follow in operations and contracts.

“This is an existential moment for creators,” Baroness Beeban Kidron writes in the report’s foreword. “Copyright is not a technical inconvenience; it is the mechanism that allows creators to earn a living and to retain control over the meaning and integrity of their work. Remove it, and you do not merely damage an industry — you dismantle the conditions under which culture itself can exist.”

You may also like

Brush up on your skills in recognizing an AI image or browse the best cameras for professionals.

In an era of artificial intelligence, real emotion and stories are poised to shape the year’s photography trends – but that data is surprisingly coming from an AI-heavy photography software. Aftershoot recently shared a survey of wedding and portrait photographers highlighting the biggest photo trends of 2026. The common thread running through all five trends? Photography will trend more towards less perfection, more human in 2026, Aftershoot says.

The survey hints that photographs that favor emotion and story – and even analog photography – will be key in 2026.

An AI-heavy software company publishing a list of trends that feel rather anti-AI feels a bit strange – but Aftershoot isn’t a tool for generating fake images from a blank canvas. The software company is known for tools like AI-assisted culling to find the best shots faster, along with AI editing profiles that learn a photographer’s signature style to speed up the editing process.

AI is still listed among the software company’s predictions for 2026, but as a behind-the-scenes tool. As Aftershoot describes it, the survey of professional photographers offers insights that “point to a future defined by emotion over perfection, substance over style, and technology working quietly behind the scenes – not replacing the artist, but supporting them.”

According to a survey of Aftershoot photographers, these are the biggest photography trends to watch for in 2026.

Emotion over perfection

One key trend for 2026 is that the photographs with the most impact will prioritize emotion over perfection. The trend predicts that factors like posing, sharpness, and other technical standards will take a backseat to emotion.

“What’s coming is more humanity and less posture,” Fran Ortiz, destination wedding photographer, told Aftershoot. “Unfocused photos that pinch you, tears that don’t get retouched, hugs that almost smell. Photography that feels like a memory already lived.”

“People are craving real expressions and real moments,” added portrait photographer Tanya Smith. “Clients want to see personality-led brands – real people with real points of view, not just pretty photos.”

Story-driven imagery

Photography has always been a way to tell a story without words, but as emotion becomes a top priority, so too does telling stories. Aftershoot predicts that story-driven photographs and documentary-style work will be a key trend for 2026.

While the trend crosses multiple genres, the documentary style is rising particularly fast in wedding photography. “The future of weddings is not about producing a flawless editorial set,” said editorial wedding photographer Joy Zamora. “It’s about transforming the couple’s story, quirks, values, and emotional world into something unforgettable.”

Analog aesthetics

Film has been gradually making a comeback – but photographers predict that analog photography is going to be even bigger in 2026. Analog photograph in part builds on the trend for imperfection, though Ortiz notes it's not a retro-looking Instagram filter but photography that feels timeless.

“Analogue is going to explode,” said wildlife photographer Paul Williams. “It’s imperfect, and it has soul. That’s why it resonates.”

AI behind-the-scenes

Importantly, Aftershoot isn’t saying that AI won’t be a key part of 2026 photography trends – but predicts that AI software will be more valuable behind-the-scenes. The company says that using AI behind-the-scenes to help with tedious tasks like culling and retouching allows the photographers to spend more time on creative storytelling.

“AI will streamline culling, editing, and color work,” predicts motherhood and maternity portrait photographer Esther Kay. “But the art remains human. The luxury look of 2026 is authenticity – real texture, real emotion, real connection.”

Personal branding and identity

The final photography trend Aftershoot photographers are predicting for 2026? Those portraits will go beyond how someone looks and be more about personal identity and personality.

“Portraits aren’t just portraits anymore – they’re identity,” Kay said. “Entrepreneurs, creators, and professionals want branding imagery that tells a story and defines their visual voice.”

You may also like

Browse the best photo editing software or the best AI culling software.

I've spent 20 years writing for websites. Most of them grew out of print magazines I used to work on. All of them try to inform, entertain and occasionally make you think. And yet right now, I'm watching Google prepare to pull the rug out from under all of us.

The tech giant is pushing something called AI Mode, and if the whispers are true, it'll soon become the default way millions search. That might sound convenient. But for people like me who create content, it's actually terrifying.

Here's what AI Mode does: instead of showing you a list of websites to click on, it simply answers your question itself. No links. No clicks. No reason to visit the sites that actually created the information Google's AI is regurgitating.

Logan Kilpatrick, a lead product manager at Google, posted on X that AI Mode would become the default "soon". Google walked that back slightly, but the direction of travel is unmistakable. After all, the search engine has already got over 100 million monthly users in AI Mode. The AI Mode tab now sits permanently on the left of search results, nudging users toward it. And they're adding more and more features that bypass normal search entirely.

So this isn't my paranoia. It's slowly, actually happening.

The brutal reality

Let me explain why this matters. Websites need visitors. Visitors come largely from Google. Visitors may then subscribe to that website, see ads on it, or buy things on it. And that money helps pay for the writers, editors, photographers, illustrators and developers who make the site worth visiting in the first place. Break that chain and the whole thing collapses.

AI Mode breaks that chain.

When Google answers a question directly, you don't click through to the source. The source gets nothing; no traffic, no ad revenue, no subscribers, no way to pay the bills. And unlike traditional search results, where publishers at least got a fighting chance at a click, AI Mode buries sources in citations nobody reads.

Search Engine Journal reports that organic traffic will "almost certainly decline" as answers stay within Google's ecosystem. They're not wrong. We're already seeing it with AI overviews; those AI-generated summaries that appear above search results. Now Google wants to make that the entire experience.

So what will the big, corporate websites – like those of national newspapers – do? Well, they'll be encouraged to pay for visibility through ads. Google made $264 billion from ads in 2024; they're not about to let that revenue stream dry up. So while they suffocate organic traffic, they'll happily sell you back the visibility you used to get for free.

What we'll lose

But what about the niche blogs that taught you to fix your vintage camera? The local news sites covering council meetings? The recipe bloggers, the tech explainers, the historians sharing primary sources? These sites can't afford to outbid major media outlets for ad space in AI mode. They'll simply vanish.

And what will we be left with? A web that's largely Google talking to itself, trained on content it no longer needs to send you to. A closed loop where one megacompany controls both the questions and the answers the world is asking.

The cruel irony is that Google's AI is only as good as the web it scrapes. If AI mode kills off the sites creating original content, what exactly will Google train its AI on in five years? Recycled AI slop? Other AI summaries? The system eats itself.

I don't want to sound melodramatic, but this feels existential. The web as we know it (chaotic, diverse, independently created) only works because there's an economic model, however imperfect, that supports it. Break that model and you break the web.

No one can't predict how fast that'll happen, of course. Things may stay the same for a while. It all may collapse overnight. I honestly have no clue. But either way, it's worth considering how this will affect you. Because we all rely on the free web. And yet soon, Google's AI may leave us with nothing but its silence.

If you'd like to help us, you can make Digital Camera World your preferred news source on Google with these five simple steps

The Evoto controversy has highlighted the uneasy relationship between the creative industry and artificial intelligence. On the one hand, AI has the power to give. Speed up workflows, distil complicated editing techniques into simple processes, and expand creative possibilities.

On the other hand, AI has the power to take. Scrape digital works, provide the consumer with the means to access the end product, and syphon clients and consumers from working creatives.

Plenty of AI platforms take from the creative industry, but Evoto has always marketed itself as a giver. Its flagship Evoto AI desktop software is a portrait-retouching application that could have easily been targeted towards social media-loving Gen Zers, dating app romantics and anyone looking for a half-decent corporate headshot (more on that later).

And yet Evoto situated itself firmly within the professional sphere, surrounding itself with working creatives and exhibiting at industry trade shows. It stood out.

This carefully curated relationship is precisely why Evoto now finds itself fighting to earn back the trust of its user base. A couple of days ago, an AI Headshot Generator appeared on Evoto’s website – and understandably, it caused a huge uproar.

It appeared to be a fully generative AI tool that required no input from a creative; feed a selfie into the generator and get a professional-looking headshot in return. Evoto removed the feature from its website, but the screenshots had already been taken. The damage was done.

An attempt to explain the situation on Facebook alongside an official statement couldn’t stop an onslaught of angry comments. Early adopter and Evoto Ambassador Sal Cincotta released a video where he questioned his relationship with the company. And the specialist photography press got hold of the story.

If that wasn’t enough, the entire situation coincided with Evoto exhibiting at the Imaging USA expo.

So where do we go from here? Well, I might be at odds with the majority of Evoto’s user base, but while straddling the line has left the AI outfit bloodied, I think we creatives have a rare opportunity to have our cake and eat it. Hear me out…

A unique opportunity for creatives

I’ve made no secret over the years that I’m not a fan of AI. In fact, I’d rather go back to the way things were. Then again, I’d rather watch the Original Trilogy. I long for the days of Nineties sports stars like Michael Jordan, Michael Schumacher and Ronaldo Nazário. And think back to a time when my phone’s only means of vying for my attention was Snake.

But my old soul knows that the artificial intelligence boom is here, and no amount of hours playing Final Fantasy III is ever going to change that.

Last year, I attended the launch of Evoto Video in New York City and had the pleasure of meeting some of the team. And while it’s very easy to dehumanize online companies as evil, money-grabbing entities, Evoto isn’t run by a System Shock AI overlord. It’s run by human beings who seemed genuinely excited by the tools they were delivering to creatives.

I say all of this for full transparency, not to belittle the gravity of Evoto’s error. I fully back the user base’s (reasoned) responses. And I’m darned angry, too. But what I won’t do is speculate.

Evoto has said: “We do not use your images or your clients’ images to train our AI models.” It’s also said: “we realize that by testing a tool that generates images from scratch, we crossed a line.” And I choose to believe it.

The user base has a choice. And it needs to be considered very carefully, because I think it could reverberate throughout the industry for years to come.

We either drop Evoto entirely or help mould it into an unlikely ally within this frightening new AI-powered reality. The latter only works if Evoto is willing to fight tooth and nail to rebuild this relationship. But we also have to be willing to cut it some slack.

If we throw Evoto to the wolves now, we risk sending out a message to similar AI outfits that a creatives-first mantra simply isn’t worth the risk – pushing them towards the ‘AI slop’ that’s threatening our very existence.

Over the past couple of days, working creatives have shown potential AI suitors their line in the sand. And I think that can only be a good thing. But over the coming weeks, months, and years, we have an opportunity to show them what can be achieved if that line is adhered to.

If Evoto is truly remorseful, it will tread very carefully from now on. It will be more open to dialogue. It will be more transparent with its intentions. And it will be more rational with its ambitions.

If the user base jumps ship now, it’ll be heading back into uncharted waters until the tide pulls it toward another AI startup. One that likely hasn’t been stung like Evoto, likely doesn’t know (or care to know) about the line we’ve drawn in the sand. And all too quickly, we could find ourselves dragged into some kind of Groundhog Day loop.

Sadly, there will always be AI companies working against the creative industry. And not a day goes by that I don’t worry about my future as a writer, photographer and content creator. But I have had to begrudgingly come to terms with the fact that AI is here to stay.

I’m not telling Evoto’s user base what it should do. I’m simply offering my perspective. And what happens next will very much depend on Evoto's continued handling of the situation and its user base’s willingness to forgive (not forget).

Right now, though, I’m minded to try working with a company that knows the industry’s boundaries, rather than starting from the ground up.

As human perspectives often do, my opinion on this matter may well change. But whether or not you agree with my sentiment, I welcome your thoughts, your criticism, your dialogue. This article doesn’t claim to hold all the answer – and that’s how you know it isn’t ‘AI-generated slop’.

You might also like...

Take a look at the best photo editing software and the best photo organizing software. Check out the latest camera news.

If I’m testing a gadget for Digital Camera World, then I usually head to the Barbican Centre in London, England. The Barbican’s sprawling maze of Brutalist architecture makes a great location for a model shoot, especially when I need to light them with an LED or test out a variety of fog machines in the location’s long, dark corridors.

When Apple lent me an iPhone 17 to review as a camera, I predictably shot at the Barbican. However, I also wanted to put the iPhone 17 through its paces in a new location, to break free of the Barbican’s familiar grey-and-brown color palette and stony textures.

I only had my model for two hours, so I decided to ask the machine mind of ChatGPT to conjure up a London-based two-hour photo walk that would give me a wide range of locations in which to shoot my model.

I hadn't really used ChatGPT before for a project like this, but I was impressed by the speed at which the app created a location-based itinerary in part of London that I’d not visited since the last century: Shoreditch.

It struggled to give me a usable map with a detailed route inscribed on it, but it did give me the following route, complete with a timetable – and each location had some suggestions on what type of subjects and shots I could expect to capture!

I checked out ChatGPT’s suggested route on Google Maps and was able to use Street View to get a better preview of the locations that it suggested I visit. Here’s the photo walk timetable, route and some potential photo subject suggestions, and even a selection of shooting tips that it suggested:

Route Summary:
Start at Boxpark Shoreditch → Walk along Redchurch Street → Head to Brick Lane → Turn into Chance & Whitby Streets → End at Arnold Circus.

Time 0:00 – 0:20
Location Boxpark Shoreditch
Style / Highlights Urban, modern lines, signage.
Photo Ideas Geometric backdrops, glass reflections

Time 0:20 – 0:45
Location Chance & Whitby.
Style / Highlights Streets Graffiti concentration, artistic vibe
Photo Ideas Edgy portraits, shade light tests, color contrast

Time 0:45 – 1:15
Location Redchurch Street
Style / Highlights Boutique storefronts, muted tones, cobblestones.
Photo Ideas Shopfront portraits, mirror reflections, lifestyle

Time 1:15 – 1:40
Location Arnold Circus
Style / Highlights Soft greenery, vintage London brick, calm atmosphere.
Photo Ideas Depth layering on staircases, natural skin tones

Time 1:40 – 2:00
Location Brick Lane
Style / Highlights Street art, graffiti, markets, lively texture.
Photo Ideas Murals as backdrops, walking shots, wide vs portrait

Shooting Tips:
Best time: Early morning or golden hour for softer light and fewer crowds. Bring a small reflector or white card for shaded shots. Test ultra-wide for murals, standard for portraits, and telephoto for compression shots. Each stop is 5–8 minutes apart — stay flexible for spontaneous moments.

So, armed with my AI-powered itinerary, I headed to London. I planned to meet my model, Maddy (@Runway.voguish), at 14:00, so I arrived an hour early to do a quick recce of ChatGPT’s proposed route. I soon realized that there was a fair bit of zig-zagging, so I re-ordered the itinerary in my iPhone 17’s Notes app.

The recce was a good idea, as it showed me that ChatGPT had indeed provided me with some fantastic locations, from the autumnal leaf-strewn Arnold Circus, to a park surrounded by beautiful Victorian red-brick mansion blocks and tenements, to the shop fronts and graffiti-sprayed walls of Brick Lane.

When I met Maddy I was completely relaxed, as I knew where I was going and what backdrops I would be shooting her against. This gave us time to fit in a coffee break in a beautiful non-franchise coffee shop, which itself provided a lifestyle photo opportunity of Maddy drinking her coffee.

I enjoyed my ChatGPT-inspired photo walk, and it certainly provided me with a much wider range of subjects and backdrops than I usually get from my regular photo walk haunts. I used ChatGPT’s suggestions as a springboard for my shoot and was happy to tweak its timetable and location suggestions once I’d recce’d the location.

It’s certainly an experiment that I can highly recommend – and as you’ll hopefully see from my shots, I captured a wide range of colorful and varied images on the iPhone 17, using its Ultra Wide 13mm, 26mm Main, and 52mm Telephoto lenses.

For a moment, when I read the headlines, I thought I was seeing an old post, but no – X’s Grok AI is being called out for undressing any photo…again. The former Twitter’s AI image editor has come under fire for answering requests like “remove her clothes” for a second time.

Last year, X added the ability for users to use an @ tag to ask the AI Grok a question and get a publicly posted response from the chatbot. In May of 2025, an investigative report showed that X’s Grok AI was not only responding to requests like “remove her clothes” but posting the resulting edited images publicly. After the report came out, the autonomous chatbot started responding to similar requests with “I can’t help with that.”

But, at the end of 2025, X added a tool that allows users to edit any photo using the AI by clicking on the “...” menu and accessing the new “edit image” option. Following the announcement, some artists announced that they would no longer be sharing their artwork on the platform for anyone to be able to just edit with AI.

The updated feature has sparked a new trend on X of generating or regenerating photos of celebrities in bikinis on the social platform – although the feature isn’t limited to only celebrity faces. Elon Musk himself even joined in by posting a Grok-generated image of himself in a bikini.

A number of users have shared posts illustrating that the chatbot has no qualms with re-imagining a person in the photo in a bikini.

While some argue the trend is designed, much like Grok itself, for humor, the Grok bikini trend is also raising concerns over not just AI, but sparking discussions on consent and privacy. A celebrity like Musk posting a bikini image of himself is one thing; X users asking to see someone in a bikini without their permission is something entirely different.

The AI’s willingness to put anyone in a bikini is not new. Along with the backlash from last May, the AI came under fire shortly after its image-generating capabilities were introduced for generating images of political candidates, including then presidential hopeful Kamala Harris in a bikini.

Along with continuing to drive artists away from the platform, the latest Grok backlash highlights the dangers of an autonomous, readily accessible AI. Since the AI’s launch, Grok has prided itself on being an AI with more lax restrictions, which were originally designed to give the AI a sense of humor.

You may also like

Browse alternative social media platforms for photographers.

I’d like to start this piece by sharing four comments that appeared below a recent Twitter/X post I made (spelling and punctuation as written…)

1. “for f#ck sake stop f#cking up perfectly good artwork with AI you HAD something already”
2. “Your photos are always hugely creative! Ignore the luddites with their pissy comments. They look AWESOME animated like this!”
3. “Why the AI sh#t instead of creative work or art”
4. “Amazing work. Ignore all the haters / bullies”

These positive and negative comments are typical reactions to my use of AI in animating photographs. They clearly show that the subject of AI and its use in ‘art’ is deeply polarising.

My hobby is toy photography. This involves posing Doctor Who action figures, lighting them atmospherically, and compositing them against photographic backgrounds that I’ve shot using my Photoshop skills. For the last few years, my humble band of around 640 @scifitoyphoto Instagram followers was happy to give my shots a few likes, and everyone seemed happy.

Then, like a new type of Doctor Who monster, generative AI began to rear its head. Back in 2023, I started using Adobe Firefly to generate the occasional alien planet or spooky graveyard background for my toy photos. As these backgrounds were secondary to the foreground figures (and often blurred to fake a shallow depth of field), no one seemed to object, and the ‘likes’ kept on coming.

Fast forward to the present day, and I’ve recently discovered NaukNauk. This is a smartphone app that’s dedicated to bringing photos of toys to life as five-second animations. I began feeding my older sci-fi toy photos into NaukNauk and was amazed at what it could produce. A still image of two Cybermen in a cemetery sprang to life as the silver robotic figures strode towards the camera, blades of grass being compressed under their metal boots. The app even preserved the original dramatic lighting, backlighting the marching Cybermen as they advanced.

Not everyone in my small social media pool was as happy as I was. One of my Instagram followers objected to my use of AI because it was "stealing from other artists". My response was that since I’d composed the toys in the original photo, lit the set, composited the elements together in Photoshop, and directed the action via text prompts, it wasn’t ‘stealing’ to then bring an image that I had created to life. My critic didn’t accept any of these points and said that it would be a “really sad unfollow for me to do”.

The negative feedback on my use of AI continued. I laughed off the occasional “this is sh*t!” comments because they were knee-jerk, and the writers didn’t appear to understand the process of collaborating with AI to add animation to an existing still image. My favourite trolling comment to date said that one of my animated action figures "looks as bland and lifeless as your face does in your profile picture". At least they’d spent some time and thought on crafting their comment, rather than defaulting to the more common and lazy “this is AI slop!” that some people regurgitate below a post.

Another common and recurring opinion was that I was damaging my creative work by using AI. “Can you please stop posting the AI slop of your great photos? It’s really diminishing the quality,” wrote one follower. I took the ‘great photos’ bit as a compliment and then had an idea to placate enemies of AI. I proposed to the comment writer that in the future, I would try to post a still image alongside an animated version. People would then need to swipe from the still to view the animation. The critical follower didn’t reply directly to this proposal, but I was pleased to see that they did ‘like’ my next post, so I took that as a seal of approval for my compromise.

I understand that the data centers that power the AI computations used to animate my toy photos are not good for the planet, and this does give me food for thought. Though only one person has used this point as a criticism. However, I can’t resist the creative scope that AI adds to my toy photography. I enjoyed creating the Doctor Who-themed images in the past, but now I can direct the ‘performance’ of the figures and make cinematic camera moves in short clips, which gets me one step closer to being a ‘director' of my favourite TV show and is immensely creatively fulfilling. Which is what a hobby should be!

Feel free to check out my toy photos (and swipe for animated versions) on Instagram - @scifitoyphoto. Compliments (and constructive criticism) welcome. Now I’m off to find a less ‘lifeless’ Twitter/X profile picture!

In 2025, AI didn’t creep in politely. It barged straight into studios, edit suites and camera bags, uninvited and fully caffeinated. If you shoot, edit or direct for a living, you didn’t just hear about AI developments this year: you lived them.

This year, what was once an “exciting technology on the horizon” became something far more immediate. It altered briefs, workflows, client expectations and, occasionally, the collective blood pressure of photographers and filmmakers.

Most strikingly, we all started to wonder what this would all mean for our careers – especially if you were in one of the photography jobs that AI is coming for.

AI image generators hit photorealism so convincingly that retouched portraits from real shoots sometimes struggled to look more polished. On the flipside, editing suites filled with tools that quietly shaved hours off masking, color correction and rough cuts. Truly, AI could giveth and taketh away.

ABOVE: Watch the Sora 2 sizzle reel

Then, just as everyone thought they’d caught their breath, OpenAI released Sora 2 in September. Suddenly we had software capable of producing physically coherent, dialog-synchronized video clips that felt like they came straight from a mid-budget Hollywood production.

Filmmakers and editors were getting a taste of what photographers had already experienced. They weren’t staring at theoretical disruption any more; it was right there on their screens, asking to be added to the workflow.

The deeper story

Yet the story of 2025 isn't simply that AI improved. Of course it improved. The deeper story was the confrontation it forced: what counts as real, what counts as ours and what creativity looks like when machines can mimic almost anything.

The launch of Sora 2 captured this tension perfectly. Technically, it was astonishing. Shots that once required entire departments could be summoned with a prompt. Movements obeyed physics, voices matched lips, shadows fell in exactly the right places. But the rollout caused immediate chaos.

The discovery that Sora’s training data included copyrighted material unless creators actively opted out provoked rightful outrage from studios and rights organizations. Directors and editors found themselves wondering if adopting the tool too soon would entangle clients in the next big intellectual-property scandal.

Then came a twist: audiences didn’t wholeheartedly embrace AI-generated video, either.

Those viral AI clips enjoyed their brief week in the sun, only to fade almost as quickly as they arrived. Filmmakers reported that when clients were offered the choice between AI-crafted footage and something shot by a human with a clear creative perspective, they still gravitated to the latter.

If anything, the excitement surrounding Sora 2 highlighted how wide the gap remains between technological capability and cultural readiness.

Industry split

Around midway through the year, it felt to me that the industry was starting to cleave into two distinct lanes.

In one lane sits the work that AI can produce cheaply, quickly and convincingly: product shoots, corporate headshots, stock-style photography, and the sort of fast-turnaround brand content that was once the backbone of many photographers’ calendars.

Much of this work evaporated almost overnight, with serious consequences for individuals' bank accounts and mortgage payments.

But in the opposite lane, something interesting happened. Creative labor that relied on emotional intelligence and spontaneity has started to flourish. Brands increasingly value the unmistakable fingerprints of human intention.

At the same time the middle ground – that comfortable space of technically competent but stylistically neutral work – has shrunk dramatically.

For photographers and filmmakers themselves, then, mixed emotions were very much the order of the day in 2025.

On one hand, AI integrated itself into every tool we used, from Lightroom to Premiere Pro. Autofocus seemed to read your mind. Color grading assistants prepared shockingly decent first passes. Retouching tasks that once drained your will to live were handled in a single brushstroke.

Even within cameras themselves, computational photography increasingly took over, with AI-powered phones like the Samsung Galaxy S24 Ultra and Google Pixel 9 leading the way.

But all this convenience came with an equal amount of unease. Subscription costs rose. Storage costs ballooned. And perhaps most painfully, skills that we'd spent years refining became optional at best, redundant at worst, almost overnight. And beneath all this ran the defining revelation of the year: the perfection paradox.

The perfection paradox

As AI became capable of generating flawless imagery at industrial scale, perfection itself lost its cultural value. Social feeds were filled with technically immaculate visuals, yet the images that gained traction were the ones that looked touched by real human hands.

Consequently, many photographers leaned into film grain, motion blur, quirky colors, accidental flare and even cameras with deliberate limitations.

Similarly, some filmmakers embraced handheld jitter, imperfect light and textures that signaled real-world presence. This wasn’t nostalgia for nostalgia’s sake, it was a deliberate response to visual saturation. When everything looks pristine, people crave the unpredictable.

High-end clients caught on to all this quickly. They didn't want work that competed with AI on technical precision; they wanted work that couldn’t be mistaken for AI at all.

Imperfection – or more accurately, human presence – became a marker of value. The more synthetic the landscape became, the more desirable the unpolished truth felt.

Conclusion

So where does that leave us heading into 2026? On the one hand, it seems undeniable that AI will remain a staple of everyday creative work. It’s too useful for battling the repetitive, fiddly, joyless parts of production and post-production to abandon. But it no longer defines the pinnacle of the craft.

Perfection has stopped being the goal. The creators who thrived in 2025 weren’t the ones who matched AI’s precision; they were the ones who focused on the parts of image-making and filmmaking that still belong wholly to humans: perspective, emotion, connection and the ability to turn a real moment into something that resonates.

In short, this year taught us that AI can imitate aesthetics but not intention. It can replicate style but not meaning. And it can generate spectacle but not feeling.

The tools changed. The economics shifted. But the heart of the craft – the spark that makes a viewer feel something – has remained firmly, defiantly human. In the end, I reckon that's the most important lesson 2025 has to offer.

The advancement of generative AI has brought with it arguments over just how much “creativity” a machine can have, but a group of researchers recently published a study that suggests generative AI image makers tend to default to just 12 general motifs or topics.

In the study, which the authors describe as “sobering for computational creativity,” two AI systems were set up to generate 100 images without human input. Despite repeating that test to generate thousands of images, the researchers found that the generations all tended to fall into one of 12 common motifs, such as a bridge, a lonely tree, an action photo, or a lighthouse scene.

The researchers described the resulting collection of images as generic-looking “visual elevator music.” The study comes from research based out of the Department of Data Analytics at Dalarna University in Sweden and the BEACON Center for the Study of Evolution in Action at Michigan State University in the US.

For the research, the group wanted to set up an AI to generate images without human input. To do that, the team gave an AI image generator a prompt to produce an image. A second AI was then asked to describe that image. The description was then fed back into the image generator to create an image based on that AI-written prompt. This went on for 100 rounds. Researchers repeated the experiment 40 times and then repeated it with four different image generators.

When the researchers looked at the results, they found that the images tended to fall into one of twelve different motifs or styles. As the researchers described it, “...the systems systematically evolved toward nearly identical semantic and visual endpoints—stormy lighthouses, urban night scenes, gothic cathedrals, and palatial interiors. Rather than exploring creative possibilities, autonomous AI loops appear to gravitate toward what could be called visual elevator music.”

“These findings are sobering for computational creativity,” the researchers wrote. “If AI systems consistently collapse toward generic outputs when operating without human intervention, this questions whether current approaches can achieve genuine machine creativity. The tendency toward ‘safe’ visual tropes suggests that maintaining creative diversity may require explicit anti-convergence mechanisms or continuous human curation.”

The consistency that the multiple image generators returned to the same 12 general topics or themes suggests “that these systems favor high-probability outputs over genuine novelty,” the researchers concluded.

The researchers warned that “widespread deployment of such systems could inadvertently homogenize visual culture.”

But, the researchers note that human culture transmission research tends to repeat common themes as well, noting themes in storytelling as well as visual arts repeated across even separate cultures and during different time periods. “The difference lies not in the presence of convergence but in the specific attractors: where humans converge on flood myths and spiral patterns shaped by embodied cognition, AI systems converge on stock photography aesthetics shaped by internet-scale training data,” the researchers explained.

Generative AI systems are trained on human-created data, the study’s authors note, wondering, “What does the convergence on common artistic motifs say about us?”

The research is available at Patterns.

You may also like

Browse the best mirrorless cameras or the best laptops for photo editing.

AI is certainly one of 2025’s buzzwords, but how much time can photographers save using automated culling and editing? According to software company Aftershoot, about 89 million hours.

The data comes from Aftershoot’s annual look at aggregated, anonymized user data, a report dubbed Snapshot 2025. The company says the data indicates photographers are facing record workloads, but automation is helping keep up.

Across 188,000 photographers, Aftershoot’s active users saved an estimated 89 million hours through culling and editing automation. That comes out to 473 hours for each photographer, which is just under a dozen 40-hour workweeks.

Aftershoot says that users processed 8.8 billion images in 2025. AI-automated culling, or the process of selecting the best shot and weeding out the duplicates and misses, looked at 6.8 billion images and recognized around 1.24 billion duplicates.

The company estimates that the time saved is about $212 million in savings on lost time – and around 17.8M kWhs.

“As photographers take on bigger workloads and shorter deadlines, AI has shifted from a nice-to-have to an essential part of running a sustainable creative business,” Harshit Dwivedi, Founder of Aftershoot, said. “Saving 89 million hours in a single year isn’t just a number – it represents time creators reclaimed for creativity, clients, and life beyond the screen.”

In 2025, Aftershoot added beta retrouching tools to the software. The company says that it has saved around 401 hours per photographer on average, with the most-used tools being those for acne, blemishes, face smoothing, stray hair removal, and teeth whitening.

Another new 2025 feature? Instant AI profiles, which take a photographer’s favorite Lightroom presets and apply them dynamically, creating consistency across images shot with different exposures, white balance, and other settings.

In the 2024 numbers, Aftershoot tallied 5.4 billion images, showing a significant jump to 8.8 billion images in 2025.

As part of the study, Aftershoot has an introductory offer for new users that discounts the first six months of Aftershoot Pro to $99 (about £75 / AU$150 / CA$137).

You may also like

Browse the best culling software or the best photo editing software.

Generative video typically starts from a text prompt, but one AI company is launching a hybrid workflow that uses both real cameras and real actors enhanced with artificial intelligence. Luma AI Ray3 Modify is an AI video generator that relies on real human performances, but then uses AI to create fantasy characters, add special effects, change the lighting, and more.

Announced on December 18, Ray3 Modify is designed to work around one of the biggest struggles of generative AI video: getting accurate human motion and emotion. Text-based generative video AI typically struggles with complex movements (as this hilarious video of an AI gymnast demonstrates).

Ray3 Modify instead takes real, camera-recorded footage and enhances it with AI. The company says that creative teams can use the platform to change the actor into a fantasy character, swap wardrobes, add makeup, alter the background, and more.

Above: see Luma AI sizzle reel

Character Reference is a tool that places a custom character over a real actor’s performance, using AI to create the character, but the actor leads the motion and facial expressions. Luma AI says that the tool is designed to lock in the likeness, customer, and identity continuity for the full shot.

The company says the AI is also designed with performance preservation, which means the actor’s original motion, timing, eye line, and delivery remain consistent despite the AI changes.

Ray3 Modify can edit original camera-shot video in other ways too, including swapping out the background, changing the lighting, or altering the weather. The AI also supports start and end frames, allowing users to create a reference image for the start of the clip and a second for the end to dictate the camera movement and the character’s changes in between.

The example clips suggest that Ray3 Modify could also be used to add special effects to low-budget shoots or “travel” to multiple shoot locations without a travel budget.

“Generative video models are incredibly expressive but also hard to control. Today, we are excited to introduce Ray3 Modify that blends the real-world with the expressivity of AI while giving full control to creatives,” Amit Jain, CEO and co-founder of Luma AI, said. “This means creative teams can capture performances with a camera and then immediately modify it to be in any location imaginable, change costumes, or even go back and reshoot the scene with AI, without recreating the physical shoot.”

The company notes a few limitations of the AI, recommending that users play with the “modify Strength Levels” tool to adjust how much (or how little) of the scene the AI reimagines. Uploading a video with an aspect ratio not yet in the AI can also degrade the quality.

Ray3 Modify builds on the previously launched Ray3 and is available to try inside of Luma AI’s Dream Machine. Free tier users can create up to eight videos in draft mode for non-commercial use. The cheapest option with 4K output for non-commercial users costs $9.99 a month, while commercial subscriptions start at $29.99 a month.

You may also like

Browse the best video cameras or the best video editing software.

Dictionary aficionado Merriam-Webster has unveiled the 2025 Word of the Year – and it has both nothing to do with photography, and everything to do with photography: Slop.

Merriam-Webster defines slop this way: “digital content of low quality that is produced usually in quantity by means of artificial intelligence.”

Slop is AI content, not photography, yet the term earning top honors for the year is enough to make anyone in any creative field grin. See, Merrim-Webster’s Word of the Year (and its runner-ups, such as “touch grass,” “tariff” and the cringe-worthy “6 7”) are based on spikes in search data and the word is often representative of the year’s current events.

Choosing a word that mocks AI as word of the year is a win for human creators in all art types, including photography. Yes, creators all year have been bemoaning generative AI. But, a word that, as the dictionary company puts it, “has the wet sound of something you don’t want to touch” becoming so widely used to describe the technology, feels indicative of the general public’s views on AI, not just creatives.

I keep coming back to a meme that I saw months back that referred to writing, but could just as aptly be used for photography. It said something along the lines of this: “If no one can be bothered to write it, why should I be bothered to read it?” If a photographer can’t be bothered to take the photo, why should anyone be bothered to look at the photo?

As Merriam-Webster puts it, “The word sends a little message to AI: when it comes to replacing human creativity, sometimes you don’t seem too superintelligent.”

I’m counting that as a win for the humans.

You may also like

Browse the best mirrorless cameras (for humans) or learn how to spot an AI-generated image.

As 2025 is coming to a close, software giant Adobe is making predictions for the biggest creative trends in 2026. As a photographer and tech journalist, I was expecting to find trends heavily influenced by AI, instead, it feels like half the predictions are anti-AI.

Adobe recently published its annual predictions for creative trends. Two of the four trends feel like an evolution of the brand’s 2025 trend report. Visuals that Engage the Senses feels a bit like a visual-focused take on the immersive art installations that Adobe highlighted in last year’s report. Similarly, the surrealism trend that Adobe called Fantastic Frontiers last year feels revamped in the 2026 trend Surreal Stillness. The Surrealism trend discusses using AI to create images that defy the laws of physics.

But, the other two of the four trends feel focused on authenticity and emotions – and as a photographer, I think those two trends are key to standing out in a world that now feels heavily focused on AI.

Connectioneering

Adobe describes the Connectioneering trend this way: “Images that Stir Emotions and Spark Connection.” Or, as Adobe Principal of Creative Services and Visual Trends Brenda Miller puts it, “People really want to feel what connects us all to the special, valuable things that make us human.”

A trend that focuses on human connections in an AI-saturated world is refreshing – but I think this trend will always be integral to photographers standing out in a world where computers can create images, not just in 2026.

Computers can’t feel emotions; they can’t feel the things that make humans human. Sure, AI can try to mimic these emotions by generating images based on a database of human-created images, but emotional photographs by creators actually capable of feeling such emotions will always have an advantage.

I was reading an article a few weeks back about a photo studio that specialized in photographing couples on a blank background and then generating a background of a popular travel destination behind them. My first thought was, why would I want a photo of myself in a place that I never got to experience?

This trend is why I think it’s important for photographers to consider not just the photos that they are getting, but the experience that clients are having in front of the camera. Are they stuck in bored poses in forced smiles? Or are they genuinely enjoying the photoshoot? I think both photos that spark emotion in the viewer and emotion in the person in front of the camera will be key to standing out in an AI world.

Local Flavors

“Local Flavors” is all about authentic experiences from around the world, Adobe says. The tech company has noticed an uptick in enterprise customers looking for authentic local and regional images, which is another positive trend for photographers moving into the coming year.

Along with the growing number of businesses looking for local photographs, Adobe also ties the trend into Gen Z and “appstinence,” or taking a break from the constant connectivity. I’m already seeing this in screen-free cameras and the resurgence of lo-fi compact cameras.

But I think the desire to take a break from constant connections applies to not just taking photos with devices that don’t have notifications but to the photos themselves. Adobe says that when brands work with local creators, the content feels more authentic and memorable.

“Embrace the culture’s authentic voice – no translators, no filters,” Adobe wrote. Like the Connectioneering trend, I think this speaks to photographers finding inspiration in their own culture and what makes them unique – and human.

Many of Adobe’s launches in 2025 have focused on AI (although not all, like the new Color Variance tool or Photoshop’s new color adjustment layer). It’s refreshing to see a list of 2026 predictions that is both a mix of AI and anti-AI.

“Some are telling stories that move us. Others are using AI as a tool to explore the world in ways we've never seen before, bringing bold ideas to life,” Adobe writes. “The real opportunity lies in embracing authenticity and innovation together — while recognizing that people remain at the heart of the creative process.”

You may also like

December is prime time for looking back on the biggest trends and predicting the next ones. Read about the most popular lenses of 2025, or browse the best photo editing software.

Another day, another AI-infused competitor charges into the photo editing software fray. This time it’s Swedish AI visual startup, Vividon, which has announced the beta launch of a promptless relighting plug-in for Adobe Photoshop. Emphasis on promptless.

The AI-infused plug-in, the outfit's first-ever release, is said to enable users to relight images by choosing from a vast library of lighting setups, designed to mimic both real-world studio and environmental lighting. The direction, intensity and mood can then be tweaked via a series of sliders.

"Lighting used to be a fixed, expensive process," explains Vividon in the demo video (embedded below).

"Rigs. Setups. But everything is about to change. Introducing Vividon – a groundbreaking, user-friendly relighting plug-in for Photoshop. No prompts. No complexity. Instead, curated lights. Truly photorealistic relighting, in any way imaginable.

"Vividon creates a clean, editable Photoshop layer ready to be used immediately. No lighting rigs. No reshoots. Massive savings. Limitless creative freedom."

This certainly sounds like a more familiar way of working for creatives than simply writing out a prompt and cycling through a series of random results.

The Stockholm-based outfit also states that adjustments are transparent, so the original image is preserved, providing what is presumably a non-destructive workflow.

This 'shoot once and decide on the light later' ethos could be useful for commercial photographers when dealing with feedback from clients. The AI startup itself says that the software, developed in collaboration with creatives, aims to “dramatically reduce the need for complex on-set setups, reshoots and heavy compositing.”

There are also plans to develop the technology further to provide more advanced controls, video platforms, and 3D pipelines. You can sign up for access to the free beta now at the Vividon website.

You might also like...

Check out the best photo organizing software. If you're looking for Adobe plug-ins, take a look at the excellent DxO Nik Collection 8. And if you want to update your hardware, here are the best MacBooks for photo editing.

As a pro photographer, I listened to Adobe’s announcement that Photoshop would be available free inside ChatGPT with a note of trepidation. But when I opened ChatGPT to try out the new integration for myself, the results weren’t at all what I expected – and that’s both good and bad.

This week, Adobe announced that some Photoshop tools would be available directly inside ChatGPT. Admittedly, I was expecting the chatbot to churn out a fully AI-generated image – but that’s not at all what’s happening here. ChatGPT is now able to help users apply select Photoshop effects, including colors and brightness, to photographs.

On the one hand, I took these photos in June and still haven’t found the time to edit them; I’m always looking for ways to speed up photo editing. But on the other hand, I’m a bit wary of AI. Here’s what I found when I tested Photoshop inside ChatGPT.

Photoshop inside ChatGPT is not generative AI

The Photoshop-ChatGPT collab is Agentic AI, where a chatbot does something for you, but it is not generative AI, where an AI creates something completely new. That means the resulting image applies edits, but does not generate new pixels. It also means that Adobe doesn’t apply Content Credentials to the image.

This is an important distinction. I don’t have an issue with AI that can help me quickly mask out a subject (provided that AI is trained on licensed content that pays the original creators). But I don’t want an AI to create for me – and that’s not what the ChatGPT Photoshop integration is doing.

Photoshop inside ChatGPT is meant to be a learning tool…

If a chatbot can handle all the editing for me, is there any reason to learn photo editing techniques? Admittedly, I was also expecting the Agentic AI to remove the learning process entirely, but I was relieved to see that that is not at all the case here.

When ChatGPT adjusted the colors of this photograph for me, the chatbot explained a bit of the difference between the saturation and vibrance sliders, for example.

I think this tool could potentially be a boost for new photographers learning the jargon of the photo editing industry, as the chatbot doesn’t just handle the edit without explanation and doesn’t entirely remove specific photographer terms either.

Another important distinction is that ChatGPT gives the user sliders to control how much (or how little) of the effect to apply.

… but ChatGPT can use the wrong terms

I got pretty excited when ChatGPT said it could apply a Kodachrome look to my photo. Imagine my disappointment, then, when ChatGPT instead delivered a slider for the tritone effect.

Kodachrome is a popular film stock; trione is an editing process that maps the highlights, mid-tones, and shadows to three different colors. The two things are not at all similar.

That, I think, is one of the downsides to trusting an agentic AI to teach the ins and outs of photography and photo editing: AI can be wrong, and mistakes aren’t exactly rare.

ChatGPT’s Photoshop is a very lite version of the photo editor

The initial launch inside ChatGPT contains only a few Photoshop features. Users can adjust color, brightness and contrast. It can apply a few special effects, like tritone and halftone dots, but that’s pretty much it, offering a very small taste of the longstanding pro photo editor.

The Photoshop ChatGPT app feels more like Photoshop Elements – and even that isn’t exactly correct, as Photoshop Elements can still do things like crop and retouch acne, while the ChatGPT collab cannot (at least not yet).

It's also important to note that while Photoshop is free, it's subject to the same limitations as the ChatGPT payment tiers. I don’t have a paid account, so I was only able to upload one image a day, and I ran out of thinking space towards the end.

Photo editing with ChatGPT wasn’t faster than doing it myself

Here’s my biggest sticking point: I tweaked the photo's colors in Lightroom far faster than ChatGPT edited the photo for me. Of course, as a pro photographer, I’m not the intended audience for this tool, but it’s a distinction worth noting.

It took some time to upload the image to ChatGPT, some time for the AI to think. And it probably took me a bit longer because there were a few things that I couldn’t figure out. I still haven’t, for example, found the ability to export a full-resolution file of the edit, though Adobe previously indicated that it was possible.

Part of this is that the feature still feels a bit glitchy. ChatGPT told me on a few occasions that the Photoshop tools were temporarily offline.

Will Agentic AI be part of the future of photo editing?

Many software brands have gone all-in on AI, but I think some AI is more ethical than others – and AI is better suited for some tasks over others.

After trying out the ChatGPT-Photoshop integration, where I see Agentic AI fitting in the industry the best is to carry out tedious, non-creative tasks. I feel like agentic AI could eventually replace things like Photoshop’s batch editing and Actions tools.

I also think Agentic AI could perhaps be an updated form of in-app help, though this is still subject to some AI inaccuracies.

Photoshop’s integration into ChatGPT isn’t going to replace pro editing – it’s not even close. But it could give new users a taste – and may indicate where future sedition automation and in-app learning may be headed. (And I’m so very glad that its not generative AI.)

You may also like…

Browse the best photo editing software. Or, learn from real photographers, not chatbots, from DCW's photography tutorials.

Learning Adobe Photoshop comes with an entire mini dictionary of photo-editing terms – but using Photoshop within ChatGPT requires knowing none of those. Photoshop now exists within ChatGPT, opening up natural language agentic photo editing, all without a Photoshop subscription.

On Wednesday, December 10, Adobe announced that Photoshop – alongside Adobe Express and Acrobat – are now available inside ChatGPT without a subscription. Importantly, those adjustments are "standard Photoshop edits and not generative edits," an Adobe spokesperson told Digital Camera World, which means the resulting image is AI edited, not AI generated.

The announcement rolls out first in the US, using ChatGPT on desktop, web, and iOS apps. Adobe indicates that the tool is expected to come to the UK and EU in the "near future," along with Android support coming soon.

The software giant already has plans for bringing additional programs into ChatGPT, adding that “this is just the beginning, we will continue to add new capabilities to our apps in ChatGPT in the coming weeks.”

The launch gives ChatGPT users the ability to ask for photo edits using natural language, such as “make this photo brighter” without knowing photography jargon like “exposure.” The Photoshop that’s integrated into the chatbot can tackle both edits to the entire picture and adjustments to only a small portion of the image.

The chatbot doesn’t stop at just making the requested adjustment though – the tool gives users sliders to tweak the results. Users also have the option to open the image in the web browser version of Photoshop for further adjustments.

ChatGPT users can call on the longstanding photo editor to edit their work by starting a prompt with “Adobe Photoshop” or by choosing Photoshop from the + icon dropdown menu.

The ChatGPT integration doesn't come with all of Photoshop's tools, but it is meant for those new to Photoshop and includes adjustments to contrast, brightness, and exposure, as well as adjustments to the background and applying creative effects. The integration also has a shortcut to continue editing inside Photoshop Web, or to open the image in the full-featured Photoshop.

I watched a demo of the tool ahead of the announcement. Adobe Principal Director Evangelist Paul Trani showed how users could ask for something specific, such as taking a photo of New York City and asking “give it a vintage look, except for the taxis.”

Trani also demonstrated how the AI could respond to less specific requests. When asked a simple “add special effects,” ChatGPT-Photoshop came up with a few different suggestions, showing a preview of the different ideas on the photo.

The ChatGPT partnership isn’t the first time Adobe has added agentic photo editing – an AI agent that edits photos for you – to Photoshop. In October, Adobe announced agentic photo editing for Photoshop, but the feature is still only a private beta and only in Photoshop Web. Integrating the longstanding editing software directly into ChatGPT brings agentic photo editing to the AI’s 800 million users.

"This builds on the foundational investments we've made in AI assistants across our own apps and allows us to meet customers where they already are as working with AI agents become the starting point for accomplishing everyday tasks. And for anyone who wants the full power of our tools it's seamless to move from ChatGPT into our native apps,” said Adobe Digital Media Chief Technology Officer Ely Greenfield.

Adobe says the feature "preserves the quality of the image." The company notes that ChatGPT performs Photoshop edits for the user, but it does not regenerate a new image. The original resolution of the image is also preserved, unlike some of the resolution limitations on some generative AI.

Because ChatGPT is applying Photoshop edits and not generating a new image, Adobe says that Content Credentials and other AI watermarking tools are not added to the image.

A perk for ChatGPT users may also be a point of contention among photographers, as Photoshop inside ChatGPT is free. Meanwhile, using generative AI inside Photoshop requires generative credits, which can drive up the cost of Photoshop (beyond the subscription alone) for heavy AI users.

Photoshop inside ChatGPT is rolling out on December 10 to the AI chatbot.

Editor's note: This post was updated with additional comments from Adobe.

You may also like

Browse the best photo editors, or browse the recap of what's new in Photoshop 2026.

Google and Alphabet CEO Sundar Pichai views Nano Banana as a tool for realizing “latent creativity.” Pichai sat down with Google DeepMind’s Logan Kilpatrick to share insight into the newly launched Gemini 3, and the interview offers insight into how the tech mogul views generative image creation.

"The other thing [Nano Banana Pro] shows me is how much latent creativity there is in the world,” Pichai said. “People are going to express themselves, and we're giving them the tools to do it the way they're thinking it in their head…We've been constrained by the tools in front of people, you may not have realized it, but we are creating more and more expressive tools and they're more and more accessible to you know more and more people.”

The idea of “latent creativity” is one of creative abilities that are dormant or not yet realized. There’s a gap in the creative process from the idea to the final product that the artist traditionally fills with tools from paintbrushes to a camera. Pichai is calling Nanao Banana Pro a tool to get from that idea to the final product, and one that is accessible to more people.

Technology has, over time, made photography more accessible – a key example being the camera, which was originally only a tool for professional photographers and chemists but now resides in most pockets via a smartphone.

The accessibility of generative AI brings a wealth of ethical debates, and training an AI using unpaid work from artists who have struggled through the creative process is just one of them. Removing the difficulty of creating can also arguably remove the incentive to learn, grow, and adapt.

Google announced Nano Banana Pro last week, shortly after the launch of Gemini 3.

You may also like

Learn how to spot an AI generated image, or browse the best photo editing software.

Another day, another AI camera. But the RocX doesn't just look different; it looks like it can also be used in conjunction with a mirrorless camera or traditional DSLR. Billed as the “World’s First AI-powered Distant View Camera,” it’s essentially a compact camera perched atop a joystick-controlled mount (RocX Handle), with the option for a larger gimbal (RocX Gimbal).

What’s interesting about the RocX Camera is that it betrays its compact form factor and is designed for long-distance photography and video. That’s not all, Farseer, the company behind it has links to an industry giant, with the RocX system said to be “Built by EX-Engineers from DJI.”

The camera, which is giving me DJI Osmo Pocket 3 vibes (but bigger), is built around a 1/2.8” sensor, has an ISO range of 100-12,800, and is said to have optical and electronic image stabilization. The rear of the camera is dominated by a large, five-inch touch-screen, which will surely feel familiar to camera phone photographers.

When attached to its mount, the device boasts a function called Auto Tracking, where it physically pans to follow the subject, with subject recognition for people, animals, and birds. There is a video of a conventional DSLR/mirrorless mounted on the RocX Gimbal, with the RocX Camera mounted in the hotshoe and the gimbal moving to track a subject under its own steam, which is certainly an interesting application. It also boasts AI-powered photography, although further details are currently thin on the ground.

As you’d expect from a device touted as a long-distance camera, it has an extremely meaty 50x zoom, which starts at 35mm and reaches a whopping 1,750mm, with an aperture of f/2.0 to f/3.2. However, it’s unclear whether the presumably fixed lens has any optical-zoom capabilities or whether the zoom is entirely digital.

The camera can also shoot 4K and fullHD video, maxing out at 4K / 30p, and 1080p / 120p, the latter allowing for smooth slow motion. Both the RocX Handle and the RocX Gimbal feature two-axis stabilisation, with the former weighing in at 500g and the latter, 1kg. The entire system has a battery life of six hours.

The Kickstarter for the RocX system launches in December, with up to 65% off Super Early Bird prices announced as follows. RocX Camera: $199. RocX Camera + Handle: $299. RocX Camera + Gimbal: $699. The message: “See You at CES 2026!” suggests that we'll see the device at the Las Vegas consumer-tech show in January.

You might also like...

If you're into AI cameras, take a look at the AI mirrorless made to ditch photography’s learning curve. This massive YouTube AI trend for '1950s Super Panavision 70' gets MILLIONS of views. And if you're not a fan of AI at all, why not wind back the clock and check out the best film cameras?

In a whirlwind of new AI announcements this week, Google’s AI has a new skill: the ability to tell users if an image is likely an AI-generated fake.

On the heels of the launch of Google Gemini 3 and Nano Banana Pro, Google announced that the Gemini app now has AI image verification. The tool allows users to upload an image to the chatbox and ask in plain language whether or not the image is AI-generated.

The tool only works with images generated with Google’s AI, so the chatbot won’t be able to say if an image was generated by another platform.

The tool works by reading the SynthID, the invisible watermarking that Google adds to its AI-generated images. Google says that users just have to upload the image in Gemini and ask: Is this AI-generated?

Demos of the feature illustrate that the tool can analyze how much of the image has a SynthID, in order to help recognize partially doctored images.

While the feature only works with SynthID for now, Google says that it has added C2PA content credentials to Nano Banana Pro, which is a standardized type of invisible AI watermark used across multiple platforms. Google says it plans to bring the ability to detect C2PA credentials into Gemini as well, which would give the chatbot the ability to detect fakes beyond just those created by Google’s AI.

Google indicates the Gemini’s ability to spot fakes is just the start. The company is planning on adding a similar feature for videos and audio. The company also says that it plans to build similar capabilities into Search.

You may also like

Read more on how to spot an AI-generated image.

Hot on the heels of the Gemini 3 announcement, Google has lifted the veil on its latest AI image generator: Nano Banana Pro. The updated AI is already inside of Photoshop, and it brings updates like the ability to generate recognizeable group photos and the ability to control the “camera.”

When Gemini 2.5 Flash Image launched, the AI was better known but the code name Google had given it, Nano Banana. Now, Google appears to be embracing that viral name, dubbing the newest image generator Nano Banana Pro.

Nano Banana Pro brings a number of improvements to the previous model. Part of Nano Banana’s fame was the ability to generate images that still look like actual people. I tried it, and it was both impressive and terrifying at the same time. Nano Banana Pro takes that a step further – it’s now capable of generating group photos of up to five people while still “maintaining the consistency and resemblance,” Google says.

The updated AI also brings more advanced “camera” controls, though of course, as an AI, there is no actual camera. Google says that Nano Banana Pro can adjust the camera angle, change the focus, and adjust the scene’s lighting.

Resolution is often a key shortcoming of AI generators, but Nano Banana Pro will be able to create graphics up to 2K and 4K in resolution, along with the ability to change the aspect ratio. The Pro model is also capable of producing infographics, an interesting capability as early AI always struggled with creating text on images.

Nano Banana Pro will also be available in Adobe Photoshop and Firefly from the first day of the new platform’s launch. Photoshop previously added the ability to choose from different AI models, which gives creators the ability to choose the AI that generates the best results for different tasks.

Inside Photoshop, Nano Banana Pro will be available to choose as a partner model when using Generative Fill.

Inside Adobe Firefly, users can use up to six reference images when working with Nano Banana Pro.

Nano Bana Pro is beginning to roll out globally today to the Gemini App by selecting "Create Images" and "Thinking" from the drop-down model menu. Free Gemini users have limited Nano Banana Pro generations before it reverts back to the old model.

Using Nano Banana Pro inside Photoshop doesn't require a separate subscription beyond the Creative Cloud, but uses generative credits. Until December 1, Adobe is allowing Creative Cloud Pro and Firefly plan subscribers unlimited generations on Nano Banana Pro to test out the new capabilities.

You may also like

Browse the best photo editing software or the best Photoshop alternatives.

How do you feel about AI right now? If you're a typical photographer, you've probably spent the last two years worried about it stealing your work. Training datasets scraping your portfolios without permission. Clients replacing you with text-to-image generators. An entire creative economy threatened by machines that learned their craft from your images.

But here's the uncomfortable truth: if the AI bubble really does burst, that might actually be worse for photographers than the current boom.

Google CEO Sundar Pichai has just told the BBC that "no company is going to be immune" if AI investment collapses—including tech giants like his own. His warning echoes the 2008 financial crisis, when seemingly distant Wall Street problems devastated working people across every sector. And so photographers should be paying very close attention.

Is this a bubble?

So how much could an AI crash affect the global economy? Well, a quick glance at the numbers shows the scale of the thing. The valuation of Alphabet, the company that owns Google, crossed $3 trillion in September. Nvidia hit $5 trillion. Analysts are tracking $1.4 trillion in deals around OpenAI. And yet no one's actually making money out of AI yet. So the comparisons to the late-1990s dotcom bubble are unavoidable.

Pichai himself acknowledged "elements of irrationality" in current AI investment, comparing the cycle to the early internet boom when "there was clearly a lot of excess investment". Federal Reserve chairman Alan Greenspan famously warned of "irrational exuberance" in 1996, years before the crash. We're hearing similar warnings now.

JP Morgan boss Jamie Dimon told the BBC that some AI money will "probably be lost." And Michael Burry—the investor who predicted the 2008 housing crash—is actively shorting the AI market. The smart money is getting nervous.

So investors could lose a lot of money, but what about the rest of us? Well, those of a certain age will remember that when the dotcom bubble burst in 2000, it wasn't just tech companies that suffered. 

The crash wiped out pension funds, triggered mass unemployment, and created a recession that touched every corner of the economy. An AI collapse would follow similar patterns, but with specific impacts on the creative sector. No one knows the future, of course, but here are my best guesses.

What might happen

The economic ripple could spread. Alphabet, Microsoft and other AI-heavy companies employ millions and support countless contractors and freelancers; including photographers who work for tech companies, shoot products for e-commerce platforms, or rely on advertising budgets from tech-sector clients. A major contraction would spread rapidly through the creative economy.

Client budgets could vanish. Corporate clients would probably respond to an AI crash by cutting spending across the board... including on human photographers. That sounds counterintuitive, but when tech sectors crash, fear spreads faster than reason, and marketing budgets get slashed indiscriminately. Believe me, I've seen it happen, time and time again.

Camera manufacturers could stumble. Canon, Nikon and Sony have all invested heavily in AI-powered features: subject detection, real-time tracking, computational photography. These R&D programs depend on the broader AI ecosystem remaining healthy and funded. A collapse would freeze innovation in camera technology just when it's become genuinely useful for working photographers.

Stock photography could crater. The companies that license your images, many of which have pivoted heavily into AI, would face existential funding crises. Several have already invested millions into generative tools. If AI investment dries up, these platforms face a brutal choice: return to traditional models they've already dismantled, or go under entirely.

The creative toolset could stagnate. Adobe's current strategy revolves almost entirely around AI. Photoshop's Generative Fill, Lightroom's AI masking, the suite of tools that now save you hours of tedious editing... all depend on continued AI investment and development. A funding collapse would halt the development of tools you've likely come to rely on.

Hopefully, of course, none of this will happen. There will be no AI crash and everything will continue to tick along. But working in the unpredictable world of media for 30 years has taught me that it's always good to be prepared. Specifically: avoid getting too much in debt, keep some savings in the bank, and have a plan B, C and D. Those are good things to do whatever happens in the wider economy.

An AI camera that doesn’t appear to exist, physically, has won an award. The aptly named AI Camera from Shenzhen-based design studio, Inspire Curve, has been recognized by the iF Design Awards, and I’ll admit that it does look like a rather slick bit of kit, albeit practically questionable. But I’m not ready to dismiss it as another AI affront to photography just yet...

When it comes to the design, I can’t help but wonder if it’s a case of form over function. Images (presumably digital renders) don’t look particularly ergonomic; the camera is essentially a cuboid with no visible grip. And whether that’s a secondary optic, an LED light, or a strobe that sits in the top-left corner of the device’s face, it looks like it’ll be a challenge to hold the camera without partially covering it.

Clearly, the Inspire Curve AI Camera isn’t targeted at photography enthusiasts, with the AI outfit even going so far as to say: “This intelligent design removes the need for professional skills, offering users an easy, creative, and fulfilling photography experience.” Now, I don’t know about you, but developing professional skills and making my own mistakes are a huge part of the fulfillment I get from photography and indeed any other skill-based hobby. But here’s the thing. It’s easy to assume that an AI camera is a generative-AI-fuelled abomination that has nothing to do with photography at all, but having researched the AI Camera further, I'm at least intrigued by its educational potential.

I can see from promotional imagery that it displays exposure settings, even when set on Full Auto Mode, so there's some educational benefit there. But what’s really piqued my interest is what’s called Intelligent Composition Assistance. Here, the AI is said to suggest the optimum composition, leaving the user to reframe the scene accordingly.

This is an interesting concept, and while it could quickly become rather restrictive, this is the kind of feature I can see mainline entry-level cameras implementing in the future. Voice Creative Guidance allows the user to describe to the camera the photo that’s in their mind’s eye, which is said to act like a photography instructor. And finally, a function called Creative Reference Community supplies users with references from other users as inspiration for a given scene.

As it stands, there appears to be no launch date, and I still have my doubts about the practicality of the design, but the aforementioned educational features do have some merit. Entry-level cameras have featured guide modes and the like for years; AI-infused 'beginner modes' certainly seem like a natural progression. The problem is that fewer mainline entry-level cameras seem to exist, but that's a whole other discussion.

You might also like...

Don't like the idea of an AI camera? Check out the best cameras for beginners. If you're on the lookout for an AI camera, Caira is an AI mirrorless that's available for pre-order. Speaking of, did this "terrifyingly accurate" AI camera just break photography as we know it?

Do you use Google Photos? Then its latest blog post, 6 new things you can do with AI in Google Photos, is well worth a read. There are some eye-catching new features in there, and two new abilities have grabbed all the headlines. First, you can now edit your images by text prompt alone, and second, you also have the option to transform them using the generative AI smarts of Nano Banana. But here's something else buried here that a lot of people seem to have missed.

Number 4 on Google's list is a new feature called personalised templates. Rolling out over the next few weeks in the USA, these promise to "use insights about you from your photo gallery to create edits unique to your hobbies and experiences". This means you'll be able to ask Google to do things like "create a name doodle personalized to me" or "create a cartoon of me and my hobbies".

Sounds harmless, right? But then consider what that actually means.

Psychological profile

When you think about it, what Google’s telling us is that the AI is going to sift through your entire photo library, analyse what you do for fun, where you go, who you spend time with, and how often, then use that data to generate 'creative' images that feel like you.

Essentially, it’s a system that builds a psychological profile from your memories. A bit like they do with murder suspects in TV dramas. All so you can auto-generate a cartoon of you on the ski slopes.

Google insists this is all about convenience: helping you 'reimagine' photos without needing to think up prompts. But the technical reality is much deeper. To make these personalised edits work, you must have features like Face Groups, Location Estimates and the Me cluster turned on. In plain English, that means giving Google's AI your facial recognition data, geolocation tracking and an understanding of your personal identity.

In short, this isn’t a one-off edit; it’s a full-spectrum analysis of your digital life. The AI wants to know whether you’re the sort of person who surfs once a summer or builds your schedule around catching waves. It needs to be able to distinguish a one-time hike, that you never wish to repeat, from a lifestyle that involves little else. That requires scanning thousands of images, recognising patterns in context, time, people and place.

And here's the clever bit: Google isn’t selling surveillance (something people are generally wary of); it’s selling creativity. The emotional pull of seeing AI generate something cute and personal like a cartoon or a doodle is exactly what makes people forget what they’re giving up.

It’s the same trade-off we’ve seen across social media and smart tech for years: sacrifice a little privacy, gain a little convenience. But this time, the trade isn’t for a targeted advert or a 'people you may know' suggestion. It’s your entire visual history being mined to feed an AI model.

What Google’s AI can learn (and how to stop it)

Remember, if an algorithm can deduce your hobbies from photos, it can also infer far more. Your income level from your home interiors. Your fitness routine from your gym selfies. Your relationships from repeated appearances. Your travel habits, even your politics if you attend relevant events. Each inference strengthens Google’s profile of you. That's all data which is incredibly valuable, both for improving its AI and, ultimately, for advertising to you.

As photographers, we instinctively understand that photos are never neutral; they reveal what we value and how we see ourselves. In this light, a photo library is a deeply personal archive, not a dataset to be mined. Yet Google is treating it as precisely that.

So if you use Google Photos and don't like the sound of this, it’s worth checking your settings. Disable Face Groups, turn off Location History, and think twice before enabling any personalised features. More broadly, ask yourself whether convenience outweighs the cost of letting an algorithm know you better than your closest friends. And the next time Google offers to 'help you be creative', remember: creativity isn’t free when it costs you your privacy.

You’d be forgiven for thinking that Camera Intelligence’s Caira, with its human name and on-board AI, was a concept dreamt up for Joaquin Phoenix sci-fi flick, Her. But fortunately or unfortunately (depending on who you ask), it’s very real indeed.

In fact, it’s the first mirrorless camera with on-board AI and is available for pre-order via Kickstarter right now. However, I’ve just finished watching a very interesting video from YouTuber Emily Lowrey from Micro Four Nerds – who's very impressed by the hardware, but has mixed feelings about the AI capabilities of this latest device from the creator of the Alice Camera.

I recommend watching her video for yourself below, as it raises some interesting and also terrifying points:

ABOVE: Is the Caira the end of photography as we know it?

I personally think the hardware is pretty neat. The Caira is an interchangeable-lens camera, which Emily explains is rather impressively built around the same Micro Four Thirds sensor found in the Panasonic Lumix GH5S.

It connects to a compatible iPhone via MagSafe (Android isn't currently supported), which is used to provide the live view interface. But here’s the thing: it’s not designed for photographers. *Wait… what?*

This is where the camera’s AI features, powered by Nano Banana, come in. As you can see in the video, Emily is able to use voice prompts to literally change the subject that she’s framed with the camera, and the results are, as she puts it, “terrifyingly, terrifyingly accurate.”

At one point, she photographs a plush toy on a shelf in her studio and gets the AI to change it to what appears to be a scarily accurate representation of a different plushie. It’s very hard to tell which one is real. It’s at this point that she poses the question: “Is this photography any more?”

To find out Emily’s thoughts, her musings on where artificially enhanced photography will end up (a scary thought to say the least), and what she thinks about the Caira’s hardware, make sure you watch the full video above.

The photographer and content creator does make a point of saying that she is impressed by the tech and that: “There had to be a camera company that did this first (...) I’m under no illusions, within the next months, and years, all of these AI features are probably going to be in big-brand cameras.”

However frightening that thought may be to the photography faithful, I think Emily might be right.

You might also like...

Are you an MFT fan? Here are the best Micro Four Thirds cameras. Perhaps you'd like to check out the best camera phones. And for more AI thoughts, in this new world of AI, photography will live and die by how transparent we photographers choose to be about the images we create.

I’ve watched entirely too many AI-generated videos that look like a security camera feed of some sort of wild animal attempting to trick-or-treat before realizing that what I was watching never actually happened. But as AI video increasingly begins to fill social media feeds, Content Authenticity (C2PA) is coming to videos.

This week, Sony announced that it is expanding C2PA protections to video. C2PA is a Content Authenticity initiative across multiple brands that uses standardized metadata tags to indicate how a photo was created. That includes tags indicating whether the image was captured by a camera or generated with generative AI.

But now, for the first time, those protections are coming to video files. Sony says its video-compatible C2PA is an industry first. The authenticity metadata also records 3D depth information, Sony says, which creates “highly accurate verification that videos were captured of actual, existing subjects.”

The feature is coming first to the Sony A1 II and A9 III, which already have access through updated firmware, as well as cinema series cameras, the Sony FX3 and FX30, and the PXW-Z300 camcorder. Support for the A7R V, A7 IV, and A1 is scheduled for after November 2025 or later, while the A7S III is expected to gain support in 2026.

Sony is also giving news organizations and broadcasters the ability to read those C2PA verifications inside Sony’s Ci Media Cloud.

Content Authenticity metadata helps verify if a photo – and now a video – came from an actual camera. I’m glad to see that camera makers are responding quickly to the explosion of AI-generated video.

But, it’s important to note that the metadata C2PA stamps work best when the platforms that share photos and videos are not only able to read the data, but also create a very clear and obvious label as to the source.

Several social media platforms support C2PA, including TikTok, Facebook, Instagram, and LinkedIn, though I would argue that such AI-generated labels aren’t always obvious enough. Facebook’s AI label on Reels, for example, requires going into the post’s menu and selecting AI info; there’s no overlay or label displayed on the video itself. How many people actually take the time to stop scrolling and dig into the menu?

Adding standardized C2PA labels to camera-created videos is a good start, and I hope more brands follow Sony’s lead, but social platforms need to play a role in making such labels not only easy to find but obvious.

You may also like

Browse the best Sony cameras or the best cameras for video.

Adobe is working on a chatbot that helps creators plan and create for social media – including integration with Lightroom and Instagram. On Tuesday, October 28, Adobe teased Project Moonlight, a chatbot built to help creators ideate, plan, and create for social media.

During a demo at the keynote for Adobe Max, Adobe shared a glimpse into the upcoming program. Project Moonlight is an AI chatbot designed around social media. The tool is designed to help creators brainstorm and create content for social media. Adobe calls it "an AI assistant that acts like a social strategist."

Lightroom integration allows creators to feed their images into the chatbot, then ask the AI for help with things like generating ideas on how to share those images on social media. In the demo, the chatbot came up with three different ideas initially. Users can then take one of those ideas and take it even further, with the chatbot recommending things like adding overlays to the images.

Project Moonlight is also tied to Firefly, including agentic photo editing, such as asking the bot to apply a Lightroom filter to the images.

But while integration with Lightroom and Firefly may be expected for an Adobe chatbot, Project Moonlight also works with Instagram. This compatibility allows the chatbot to make recommendations that stay on brand with your previous content, as well as offer advice and insight based on how past posts have performed.

As a Sneak, Project Moonlight isn’t available yet, but something Adobe has in the works. The software giant says Project Moonlight is “coming soon.”

• See all the news from Adobe Max 2025

You may also like

Browse the best photo editors or the best cameras for Instagram.