Twenty years ago, seeing someone on screen meant they were probably there. Today, it doesn't.
AI can generate a face in seconds. Clone a voice from 15 seconds of audio. Produce video that looks identical to a real person sitting in front of a real camera.
Every authentication system is still asking: "Who are you?" Almost none asks: "Are you continuously there?"
We call this the Continuity Problem — and we think it's the missing primitive of the AI era.
What broke Identity and presence used to be coupled. If you could verify someone's identity, you could reasonably infer they were there. AI decoupled them. The identity signal is intact. The presence signal is gone.
Why it matters now • AI agents with signing authority need to prove they're the same entity across time • Remote work is permanent — we make critical decisions over unverifiable video • Deepfakes are moving to real-time
The "seeing is believing" era is ending.
What we're doing We're a small research group called The Continuity Lab. We've run 576 experiments investigating whether continuity — the property that the same entity has been continuously present — can be made as verifiable as a cryptographic signature.
All our specifications are open. Our code is public. We don't have all the answers — and we publish what we don't know alongside what we've found.
This isn't a product. It's a research question. And we think it's one of the most important ones of the next decade.
Research by The Continuity Lab. myshape.com/continuity
#AI #Cybersecurity #DigitalIdentity #MyShape #TechResearch