Adobe's annual MAX Sneaks session, hosted this year by actress and comedian Jessica Williams, provided a preview of experimental technologies developed by the company's research and engineering teams.
 The event allows Adobe staff to present prototype concepts and gather feedback from the creative community, with many innovations more than likely being integrated into future products. 
This year's demonstrations focused on AI-driven tools, showcasing new methods for editing light and sound, as well as technology for converting 2D still photographs into 3D environments.
When is a photograph no longer a photograph?
To put it simply, the advancements in the prototype software at MAX Sneaks are 'scary good.'
We are on the verge of a future where it will be incredibly easy to remove anything from a photograph or even a video in mere seconds. This isn't just about removing a stray car in the background; we are talking about impossibly complex scenarios, like making garbage cans in the background of video footage magically vanish as a person walks right past them, or a boat with its accompanying waves.
 
From Project Frame Forward at Adobe MAX Sneaks. Image: Tim Levy
On top of this, it's ridiculously easy to add anything into a photo or video and have the results look completely realistic. Want to add a puddle? Done. It will even generate a perfect reflection of anything that walks by it.
Need to change a surface or background in a product shot? That’s literally just a few button presses away. You can take one single photograph of a product and, in a few seconds, transform it into what looks like a massive, big-budget photo shoot, complete with multiple angled options.
 
From Project Scene It at Adobe MAX Sneaks. Image: Tim Levy
How about photographing an outdoor wedding and the couple are standing under an arbor with mottled lighting on their faces? No problem - this can be fixed in seconds. 
Oh no – your image is really badly lit because you are not a professional photographer with a deep understanding of lighting? Nor do you have the budget to spend $10,000 on studio lights? No problem – just change the actual light source, the intensity and the direction within seconds.
 
From Project New Depths at Adobe MAX Sneaks. Image: Tim Levy
The prototype projects
For photography and 3D work, several prototypes were shown:
Project Light Touch: A generative AI tool for adjusting light sources within a photograph after it has been taken.
Project Trace Erase: A tool using diffusion transformer models to remove objects from an image, which also removes their related shadows and reflections.
Project Surface Swap: An AI-powered tool to select and replace materials or textures in a photo (such as upholstery or flooring) while maintaining the original's lighting and perspective.
Project New Depths: Editing tools designed for "radiance fields" (3D photos) to adjust colour, shape, and composition in a 3D space.
Project Scene It: A tool that uses Image-to-3D and 3D-to-Image technology to build 3D scenes, allowing users to tag individual objects with reference images and position them in 3D.
For video, animation, and audio, the following projects were demonstrated:
Project Frame Forward: This tool applies edits across an entire video clip based on a single annotated frame or a text prompt, intended to reduce frame-by-frame editing. Basically you take a frame from the video clip - drop it into Photoshop and erase the unwanted subject. Then bring it back into Frame Forward and AI deletes that subject from the whole video clip.
Project Motion Map: An AI tool that analyses static vector graphics to automatically generate animations without manual keyframing.
Project Sound Stager: This system analyses a video's visuals and pacing to automatically generate layered soundscapes. It also allows users to modify the mix through conversational AI prompts.
Project Clean Take: This was incredible and a super useful tool for any run and gun videographers who are working on location, or with talent that are not professional speakers. It's an AI tool for editing dialogue, designed to correct mispronunciations (you can change via text), isolate specific voices, and remove background noise.
You can read more about Sneaks on the Adobe Max blog here.
Or you can watch the Adobe Sneaks full demonstration below.


