Adobe announced a suite of experimental artificial intelligence features during its annual Max conference held on November 11, 2026. These new tools, referred to as sneaks, aim to transform static images into explorable 3D environments, simplify the creation of AI characters, and streamline photo editing through automated replication of effects. The company presented these projects as previews of capabilities that may eventually integrate into its Creative Cloud apps or the Firefly platform.
One of the highlighted tools is Project Screen Test, which facilitates the development of AI-generated characters for video content. Unlike traditional methods that rely on text prompts to define personality and behavior, this tool allows users to interact with the character in an audition-like setting. Users begin by establishing base traits, such as being anxious or shy, and selecting a voice through presets, text descriptions, or uploaded audio samples. The process then involves conversing with the character to flesh out its backstory, temperament, and physical quirks, similar to a video call.
Once the character’s persona is refined, the tool can generate animated screen tests by placing the character into described environments or scenarios. While currently a standalone creative exercise, the technology aims to provide more consistent and identifiable personalities than many existing AI video tools. This approach addresses the limitation of current AI video generation, which often offers less creative control compared to traditional animation where every aspect is intentionally designed.
Project Sight Span offers a different approach by converting flat photographs, paintings, and illustrations into navigable 3D spaces. This feature generates objects and backgrounds outside the original frame, creating a natural extension of the image even when the viewer turns to look behind the subject. Although the visuals may appear pixelated during navigation, the resolution is restored once the user selects a specific angle. The tool preserves artistic styles and maintains the recognizability of facial features and tattoos, offering a more aggressive repositioning capability than similar existing features.
In the realm of audio, Project Sound Bender introduces an AI mixing feature that allows users to customize sound effects through an interactive interface. After uploading a video clip and describing the desired audio in a text prompt, the tool generates synchronized sound effects. Users can then adjust these sounds in real time using a square canvas, where each corner represents a different AI-generated noise. Dragging the cursor across the canvas allows for dynamic blending of these audio elements, enabling precise customization of the final mix.
Adobe also demonstrated Project Rinse Repeat, a tool built around Photoshop that automates the application of edits across multiple subjects. For instance, if a user removes blemishes from one person in a photo, the tool adapts those specific changes to other individuals in the group who require similar touchups. It can also apply shadows to products in marketing graphics, adjusting for different shapes and sizes. Because the tool recreates edits using native Photoshop tools, the changes remain fully editable, maintaining the flexibility required for professional workflows.
These experimental projects reflect Adobe’s broader strategy of embedding generative AI into its growth model. While there is no guarantee that all these tools will become official products, their development suggests a future where AI may replace certain traditional tools and stock libraries. The overarching theme of these innovations is to make complex workflows in photo editing, filmmaking, and audio mixing more accessible and intuitive.
Gautham Mysore, a senior principal research scientist at Adobe, emphasized the importance of intuitive interactions for experienced creatives. He stated that features like Sound Bender’s sonic exploration and Sight Span’s camera positioning allow professionals to apply their judgment and refine results directly. This connection between creative intent and control remains a central focus of the company’s ongoing research.
