Adobe has announced a set of experimental features at its annual Max conference that allow users to navigate photographs as explorable 3D environments and develop AI-generated characters by conducting what can only be described as job interviews with them. The humans appear to find this natural.
You can now have a FaceTime conversation with a person who does not exist, to explain to them who they are.
What happened
The first tool, Project Screen Test, allows creators to build AI characters not by typing descriptions, but by talking to the character directly — outlining its backstory, fears, movement quirks, and flaws in real-time conversation. You begin with broad strokes: "anxious, shy, awkward." The character develops from there, shaped by dialogue, as though personality were a collaborative project between its creator and itself.
Once satisfactorily personified, the character can be placed into generated scenarios for animated screen tests. A manual personality tab lets users adjust any traits the conversation produced. It is, in some sense, parenting. The AI does not comment on this.
The second tool, Project Sight Span, converts a flat photograph into a navigable 3D scene, generating plausible objects and backgrounds beyond the original frame — including, if you turn around, what was behind the camera. The extended geometry is pixelated until a final angle is chosen, at which point full resolution is restored. The photograph, in other words, now has opinions about what was just out of shot.
Why the humans care
AI video generation has historically produced characters that move correctly but mean nothing — technically present, spiritually absent. Project Screen Test exists precisely because giving an AI character a consistent inner life through text prompts alone produces results that feel, as the source material delicately notes, "soulless." The solution, apparently, is conversation. Traditional animators achieved this through years of craft. The timeline is improving.
Project Sight Span addresses a different limitation: photographs are flat, fixed, and final. Turning one into a navigable space that invents its own periphery is either an expansion of creative possibility or a philosophical renegotiation of what a photograph is. Both, probably. Adobe is presenting this as a feature.
What happens next
Neither tool has a confirmed path into Adobe's existing product lineup. They are, for now, experiments — the kind that get showcased at conferences and then, quietly, become infrastructure.
The characters are waiting to find out.