Enabling a new model for healthcare with AI co-clinician
DeepMind Blog · 2026-04-30
(No summary yet for this item — extraction summaries are still backfilling.)
Appears in
Extraction
Topics: medical-aiai-co-cliniciantelemedicinemultimodal-aihealthcare-ai
Claims
- Google DeepMind is launching an AI co-clinician research initiative built on a 'triadic care' model where AI assists patients under physician supervision.
- In 98 realistic primary care queries, AI co-clinician recorded zero critical errors in 97 cases, outperforming two widely-used physician AI tools.
- In simulated telemedical consultations, the AI performed at or above primary care physician level in 68 of 140 assessed consultation skill areas.
- Expert physicians still outperformed the AI overall, especially in identifying red flags and guiding critical physical examinations.
- A dual-agent architecture (Planner + Talker) is used to enforce safe clinical boundaries in real-time patient-facing settings.
Key quotes
We hypothesize that the next evolution of healthcare delivery will entail 'triadic care' where AI agents can help patients in their care journeys under the clinical authority of their physician.
We assessed over 140 aspects of consultation skill and found that expert physicians performed better than the AI system overall, particularly in identifying 'red flags' and guiding critical physical examinations.
AI co-clinician performed at a level comparable to or exceeding primary care physicians (PCPs) in 68 of the 140 assessed areas.