Adobe has introduced an array of experimental AI innovations, known as “Sneaks,” designed to simplify creative tasks across video, photography, and audio. These early-stage projects showcase Adobe’s vision for the future of content creation, where complex edits can be performed with minimal effort and advanced automation.

According to Adobe, the new AI models can instantly apply a single edit across an entire video sequence, reshape lighting in photos, and even correct spoken errors in audio clips. While these tools are still in the experimental phase and not yet available to users, Adobe’s track record suggests many could eventually make their way into Creative Cloud applications. Several existing features, such as Photoshop’s “Harmonize” and “Distraction Removal,” originated as Sneaks before becoming mainstream tools.

Highlights of Adobe Sneaks

Project Frame Forward

Project Frame Forward eliminates the need for boring frame-by-frame masking. Users can remove or insert objects, or even people, from a video with a single action. The system uses text prompts like “add a rabbit” or “add a chair” to introduce new elements while maintaining realism throughout the entire clip. The result resembles an advanced, video-wide version of Photoshop’s “Content-Aware Fill.”

Project Light Touch

The project light touch tool focuses on lighting manipulation in photographs. With AI-driven controls, users can adjust the direction, intensity, and colour of light to reshape a scene dramatically. It can simulate turning lights on or off, dragging a light across a room, or redirecting a photo from daylight to a nighttime ambience. Light Touch also generates coloured and glowing effects, giving photographers more control over mood and atmosphere.

Project Clean Take

Project Clean Take brings sophisticated AI editing to audio production. It corrects mispronounced words, adjusts vocal tone, whether more energetic or calmer, and enables word-level changes without requiring a re-recording session. The tool can also isolate voices from background noise, offering clearer and more polished audio with minimal manual intervention.

Additional AI Experiments Previewed

Adobe also revealed a series of other forward-looking Sneaks:

• Project Surface Swap: Recognizes materials in photos and replaces them with new textures, such as swapping fabric on a wall or converting wooden flooring to a marble finish, while maintaining natural lighting and perspective.
• Project Trace Erase: Removes unwanted objects along with their shadows, reflections, and distortions, producing realistic results with almost no manual cleanup.
• Project New Depths: Introduces 3D manipulation into photo editing, enabling creators to adjust colours, shapes, and composition as though sculpting within three-dimensional space.
• Project Motion Map: Automatically animates static images, such as vector graphics, by analysing their structure and generating natural, expressive motion, without the need for keyframing.
• Project Sound Stager: Uses AI to create multilayered soundscapes based on a video’s visual cues and emotional tone. Users can refine the final mix through an interactive chat interface with an AI “sound designer.”

Adobe’s latest Sneaks highlight the company’s ongoing push to integrate generative AI into every layer of the creative process. Although these innovations are still in development, Adobe’s history suggests that many of them could become standard features in future versions of its applications. If adopted, they could significantly reshape workflows for designers, video editors, photographers, and audio engineers worldwide.