Photography is subjective. At least, that is what we tell ourselves when nobody likes a picture we were particularly proud of. But if you have ever wondered what an entirely impartial, unfeeling robot thinks of your pictures, Adobe’s Project Indigo camera app is ready to offer an opinion.
Adobe has added an experimental AI Playground to Project Indigo, its iPhone camera app designed around manual controls, computational photography, and a more natural, SLR-style image. One of its headline features is Photo Guidance, which analyzes a photo and explains what worked, what did not, and how the image could be improved.
It also features a Pixel-esque mode that can suggest how to reshoot the scene while you are still standing in front of it, potentially recommending a different position, framing, lens, or approach to the light. Adobe says it is refining the system so that it understands the limitations of individual phones and does not suggest impossible changes, such as adjusting the aperture on a fixed-aperture smartphone lens.
In theory, this sounds genuinely useful. Photography books, courses, and camera clubs have been offering constructive criticism for decades, so putting a virtual tutor inside a camera app makes a certain amount of sense. Feedback while you are still at the location could be far more helpful and genuine than an AI editor “fixing” everything later.
However, yes, there is something slightly bleak about taking a photograph and immediately asking an algorithm whether it approves. Photography is not a science, and some of the most memorable images succeed because they ignore conventional rules. An AI trained to spot familiar ideas about composition, lighting, and subject separation could easily reward technically tidy photographs while discouraging anything genuinely original.
Project Indigo’s Photo Critique analyzes a candid portrait, identifying both the strengths of the image and areas that could be improved.
Image credit: Adobe
A second Photo Guidance option offers practical advice for reshooting the image, alongside suggested adjustments that could be made afterward in an editing app.
Image credit: Adobe
More Gen AI
Adobe is also adding other AI Playground features to the app, including object removal, simulated shallow depth of field, lighting adjustments, artistic styles, and a Custom Edit box for text prompts. Adobe is initially using Google’s cloud-based Nano Banana model, with generated results limited to roughly 2K pixels per side and an internet connection required.
Michele Marieschi’s 1741 engraving of Santa Maria della Salute in Venice, used as the starting point for an AI-generated transformation.
Image credit: Adobe
Adobe asked the AI to convert Marieschi’s engraving into a color photograph without changing its original composition.
Image credit: Adobe
The original Project Indigo photograph shows a cyclist riding along a road enveloped in thick fog.
Image credit: Adobe
The same photograph after applying AI Playground’s Golden Hour Lighting style, replacing the cool gray fog with warm, low sunlight.
Image credit: Adobe
The original photograph shows a portrait taken inside a crowded art gallery, with visitors filling much of the background.
Image credit: Adobe
AI Playground’s Distractor Removal tool removes background visitors while attempting to preserve the main subject and gallery surroundings.
Image credit: Adobe
For now, access is free, requires no Adobe sign-in, and is being offered to only a small percentage of Project Indigo users for a few weeks. The feature appears after updating to version 1.1, but only works with photographs captured inside Indigo.
Adobe says the experiment could be expanded, extended, or eventually turned into a paid feature. Until then, a lucky group of iPhone photographers can discover whether AI thinks their masterpiece is genuinely inspired, or whether they should take two steps left and try again.
