The Lab
Chief Health
AI Lab.
Applied R&D partnerships and published research.
Applied R&D partnerships and published research.
Should your health system deploy ambient AI scribes — where, for whom, and what should you measure? A living review of the published evidence on documentation time, burnout, efficiency, and workflow fit.
Read the Atlas →What does the evidence predict for your context? Explore how the current research applies to your product choice, baseline burnout rate, implementation approach, and practice setting.
Open the simulator →Try MedGemma running locally on a physician's MacBook Pro in Boston — no cloud, no API, no data leaving the machine.
Try the demo →Record or upload audio; Whisper transcribes it and MedGemma generates a structured clinical note — both running locally on a laptop, with nothing stored.
Try the demo →What happens when you ground MedGemma in 2,800+ pages of trusted clinical protocols on-device? It cites its sources, stays in scope, and knows when to escalate.
Try the demo →Local-first health AI in your pocket — Google's Gemma 4 E2B running entirely on iPhone, with on-device health state extraction, scribe, and memory.
Join the TestFlight ↗Federal grant, foundation submission, cross-institutional partnership, or internal AI business case — twenty minutes is enough to know if we're a fit.