Behaviour data as a health signal — and why "privacy first" isn’t a platitude here.
Category · AI & Agents
Behaviour as signal.
Digital phenotyping is the continuous, passive measurement of human behaviour through digital devices — usage patterns, movement, sleep, interaction rhythms — and its evaluation in models that estimate mental and physical states.
Health research has established that state changes often show up earlier in behaviour than in self-reports.
Why it has to be privacy-first.
Digital-phenotyping signals are highly sensitive. Every serious product in this space therefore runs on-device: raw data never leaves the device, only aggregated, anonymised signals are transmitted — if at all. That isn’t a feature, it’s the entry ticket.
