Adobe Ditches Its Own AI for Google's Nano Banana in New Camera App Update
Adobe has made an unusual strategic choice: it's powering the generative AI features in its new Project Indigo camera app update with Google's Nano Banana model rather than its own Firefly AI. The version 1.1 update, released on July 20, 2026, adds an AI Playground suite that lets iPhone users apply generative filters, remove objects from photos, and get AI-powered critiques of their shots, all running on Google's image model instead of Adobe's own technology.
This decision is striking because Adobe has spent the past two years positioning Firefly as the responsible, commercially safe alternative for creative generative AI. The company has heavily marketed Firefly's training data provenance and support for Content Credentials, a metadata standard that labels AI-generated content. Yet here is Adobe, integrating a competitor's model into a first-party app, suggesting either that Firefly's image editing capabilities aren't yet where the Indigo team needs them, or that Adobe is willing to test against its own stack to see what users actually prefer.
What Does the AI Playground Actually Do?
The AI Playground suite in Project Indigo's update includes four distinct sections designed to handle different editing tasks. Each section contains multiple buttons or toggles that trigger Nano Banana with carefully engineered prompts. Pressing a button a second time typically yields a slightly different result, giving users some creative variation without requiring manual prompt writing.
- Styles: Applies preset generative filters that transform photos into illustrations or change lighting conditions to match different moods or artistic styles.
- Object Editing: Removes people and background distractions from photos and fills the gaps, similar to Google's Magic Eraser tool.
- Photo Guidance: Uses AI to critique your shots and suggest reshoots, offering feedback on composition, lighting, or other technical aspects.
- Custom Edit: Provides a free-form prompt box that lets users describe edits in detail, including stacking changes on top of a Styles preset for more complex workflows.
Access to the AI Playground is free during the test phase, with no sign-on required. However, Adobe is rolling out the feature to only a small percentage of Indigo users over the next few weeks to gather feedback before a wider release.
Why Would Adobe Choose Google's Model Over Its Own?
The decision to launch on Nano Banana rather than Firefly is the most revealing part of this story. Marc Levoy, the Adobe fellow leading the project, framed the choice as an experiment.
"Our goal is to learn how people use GenAI-powered editing. Depending on its success we may extend the experiment, or expand the user pool, or run follow-on experiments," said Marc Levoy.
Marc Levoy, Adobe Fellow
This suggests Adobe is running a controlled A/B test on a competitor's model to gather hard data on what edits users actually want before committing Firefly engineering resources to match those capabilities. If Firefly's next image release closes the gap, expect the model swap to be quiet, and expect Content Credentials support to arrive at the same time. For now, Adobe is shipping a camera app whose most interesting features are powered by Google.
Google has been aggressive about pushing Nano Banana into third-party surfaces, and this partnership gives the search giant another high-profile integration. For Adobe, the move puts the company on the other side of a debate it usually leads. Firefly launched with heavy emphasis on training data provenance and Content Credentials as the answer to AI transparency questions. Yet Indigo's AI Playground currently ships without Content Credentials on generated edits.
How to Understand the Transparency Gap
- Current Status: Indigo's AI Playground generates edits without Content Credentials metadata, which would normally label AI-generated content for transparency.
- Watermarking: Nano Banana already stamps outputs with Google's invisible SynthID watermark, providing some form of AI detection even without visible metadata.
- Adobe's Plan: The Indigo team is working to add Content Credentials support, though no timeline has been announced for when this will be available.
- Shared Responsibility: Marc Levoy framed the transparency gap as a shared responsibility rather than a technical fix, stating that whether the edits you apply are "fair" or "misleading" depend on what you do, who you share the images with, and how you present them.
This framing reads defensively coming from Adobe, which has spent years positioning Content Credentials as the definitive answer to exactly this question. The line suggests that responsibility for transparency lies partly with users, not just with the tools they use.
What's the Commercial Strategy Here?
The commercial logic is straightforward. If the AI Playground feature proves popular, Adobe plans to eventually offer a paid version. Adobe already sells generative credits inside Photoshop and Lightroom; a mobile capture-to-edit loop with paid AI passes fits the same model. Indigo gives Adobe a direct-to-consumer surface that doesn't require a Creative Cloud subscription to test, lowering the barrier to entry for users who want to experiment with AI-powered editing.
Project Indigo itself launched in 2025 as a computational photography app aimed at recreating an SLR-like look on the iPhone, with manual controls for focus, shutter speed, ISO, and white balance. The AI Playground update represents a sharp pivot from that original pitch. The Verge described the original app as a "camera app built by camera nerds for camera nerds," emphasizing natural output and minimal computational trickery. Now Adobe is dropping generative fill, style transfer, and text-prompted editing into that same shell, fundamentally changing what the app is.
The Verge
Marc Levoy hinted at this direction from the start, describing Indigo's debut as "the beginning of a journey" toward integrated mobile capture and AI editing. The AI Playground update fulfills that promise, but in a way that raises questions about Adobe's confidence in its own generative AI technology. For now, Adobe is shipping a camera app whose most interesting features are powered by Google, and that strategic choice tells a more revealing story than the features themselves.