Adobe has presented 10 experimental AI projects at its MAX 2026 conference, offering a look at how artificial intelligence could reshape photography, video production, sound editing and graphic design.
Showcased during the company’s annual MAX Sneaks event on November 11, the projects explore ways to make creative workflows more flexible while giving users greater control over the editing process. From turning photographs into navigable 3D environments to building interactive AI characters, the demonstrations go beyond conventional content generation.
These projects are currently experimental, and Adobe has not confirmed a general release schedule for most of them.
New AI Experiments Focus on Video Production
Several demonstrations explore how AI could simplify complex video editing tasks.
Project Scene Sweep is designed to recognise subjects, locations and scenes across multiple shots. This could allow editors to apply consistent changes across different camera angles without manually repeating each adjustment.
Project Screen Test takes a different approach to character creation. It allows users to define a fictional character’s appearance, voice and personality before interacting with it through conversations.
Another concept, Project Live Set, focuses on making changes during video playback. Creators can experiment with subjects, backgrounds and camera perspectives while working on a scene.
Project On Cue explores how visual timing signals can control movements and effects in AI-generated videos. These signals can be created manually or taken from existing footage and music.
More Flexible Audio Editing With AI
Adobe’s Project Sound Bender focuses on making individual sound elements easier to manipulate.
The tool allows users to describe a sound and separate it from a larger audio mix. Once isolated, the element can be modified or transformed without rebuilding the entire soundtrack.
It also explores generating sound variations from written descriptions and coordinating audio changes with video content.
Turning Photographs Into Interactive 3D Scenes
One of the more distinctive experiments is Project Sight Span. It explores how a single photograph could become a navigable visual environment.
Instead of remaining limited to the original camera perspective, users can explore alternative viewpoints and create new compositions from the same image.
Adobe also demonstrated Project Deep Brush, which uses image understanding to make editing more responsive to objects, shapes and lighting. Its technology can follow object contours, adjust colours while preserving natural shadows and introduce lighting changes.
Project Hands On takes another approach by placing interactive editing controls directly on an image. It identifies individual elements and allows creators to manipulate them through on-canvas controls.
Making Repetitive Design Work Easier
Adobe’s remaining experiments focus on improving editing efficiency and graphic design workflows.
Project Rinse Repeat allows users to transfer visual adjustments from one part of an image to another. These changes are recreated using editable Photoshop tools, giving creators the flexibility to refine them later.
Project Link Up explores how AI can interpret a designer’s gestures. When an object is moved into a composition, the system can adjust its scale, perspective and appearance to better match the surrounding design. The resulting vector elements remain editable.
Selected Experiments Reach Adobe Labs
Adobe is also bringing some of its experimental technologies closer to creators through Adobe Labs.
Its new Labs Playground provides hands-on access to early versions of Project Deep Brush and Project Hands On. Users can also preview Project Custom Filters, another capability being explored for Photoshop.
However, demonstrations at Adobe Sneaks do not necessarily become commercial products. The company has not confirmed when, or whether, all 10 projects will reach its mainstream creative applications.
The latest showcase indicates that Adobe is exploring a broader role for AI in creative software. Rather than focusing only on generating new content, these experiments aim to help creators control scenes, modify individual elements and reduce repetitive work while retaining the ability to make detailed creative decisions.
