On November 11, 2026, Adobe demonstrated four prototype artificial‑intelligence tools that could reshape how creators work with characters, images, sound and photo retouching. The previews, presented before the company’s MAX Sneaks session, were identified as Project Screen Test, Project Sight Span, Project Sound Bender and Project Rinse Repeat. While Adobe has not pledged to ship any of the concepts, the demonstrations suggest a future where complex creative tasks become more interactive without relying solely on textual prompts.
Screen Test and Sight Span: New Ways to Build Characters and 3‑D Spaces
Project Screen Test begins by letting users define a character through a blend of personality descriptors, vocal samples and optional written notes. The system then conducts a conversational audition, during which creators can flesh out the character’s backstory, temperament, objectives, fears and even subtle movement habits. A dedicated “Personality” tab offers manual fine‑tuning of the generated traits, giving artists a degree of direct control over the AI’s output.
After the character is shaped, the tool can place it inside a user‑described setting and generate a short animated screen test. In the demonstration, the resulting characters displayed a more coherent and recognizable personality than many existing AI‑generated animated figures, according to The Verge’s coverage. However, Screen Test currently exists as a stand‑alone exercise, and its integration with Adobe’s broader video‑editing suite remains uncertain.
Project Sight Span tackles a different challenge: turning a flat image—whether a photograph, painting or illustration—into a navigable three‑dimensional environment. The AI extrapolates objects and background elements beyond the original frame, allowing viewers to shift perspective as if they were moving around a virtual set. In the test run, tattoos and facial features retained their recognizability when the viewpoint changed, and the system managed to preserve the original artistic style.
Despite these promising aspects, the navigable space displayed noticeable pixelation and a lack of definition during movement. Resolution was restored only after a specific viewing angle was selected, indicating that the technology still needs refinement before it can serve production‑level demands.
Sound Bender and Rinse Repeat: Interactive Audio Mixing and Smart Photo Editing
Project Sound Bender offers a spatial approach to sound‑effect generation. Users supply a text description, and the AI creates corresponding audio clips that are placed in a square canvas, each corner representing a distinct sound. By moving the cursor across the canvas, creators can blend the four sounds in real time, and each corner can be individually customized. This design encourages exploration of multiple sonic combinations rather than forcing a single, static choice.
Project Rinse Repeat focuses on repetitive editing tasks within a single photograph. In one scenario, the tool transferred blemish‑removal edits made to one person across other subjects in a group photo, automatically adapting the changes to each individual’s features. A second demonstration added a shadow to a single product and then reproduced that shadow across additional products of varying shapes and sizes. Built on Photoshop, Rinse Repeat applies edits using native Photoshop tools, ensuring that the resulting changes remain fully editable—a critical factor for professional workflows that demand both automation and precise control.
Creative Intent, Control and the Road Ahead
Adobe research scientist Gautham Mysore highlighted that the interaction models behind the four projects aim to give seasoned creatives a tighter grip on the creative process. He cited Sound Bender’s real‑time sonic blending and Sight Span’s camera‑position adjustments as examples of how professionals could apply their judgment while the AI handles the heavy lifting. Mysore emphasized that preserving the link between creative intent and hands‑on control is a central focus of Adobe’s exploration.
The official MAX Sneaks session, listed for 5:30 p.m. EST on November 11, was positioned as a “first‑hand look at work underway in Adobe Labs.” The timing of the session, later in the day than the morning reveal of the projects, underscored Adobe’s strategy of teasing experimental work before committing to a product roadmap.
For practitioners in the creative industry, the key question remains whether the level of editable precision demonstrated in these prototypes will translate into marketable features. Screen Test still lacks a clear pathway into Adobe’s existing video‑editing ecosystem, and Sight Span’s visual fidelity issues during navigation raise doubts about its suitability for high‑end production. Nonetheless, the experiments illustrate a clear intent to embed AI more deeply into creative pipelines, moving beyond simple text‑to‑output models toward tools that respond to nuanced, real‑time user interaction.
As Adobe continues to iterate on these concepts, the industry will be watching to see if the prototypes evolve into polished, integrated components of Creative Cloud or Firefly. If they do, creators could soon benefit from AI‑driven workflows that blend the speed of automation with the nuanced control long demanded by professional artists and editors.






Be First to Comment