{"data":{"event":{"id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","slug":"amie-osce-diagnostic-dialogue","title":"AMIE outperforms primary-care physicians on a text OSCE study","short_summary":"Google’s AMIE diagnostic-dialogue system beat 20 primary-care physicians on most specialist and patient-actor axes in a randomized text-chat OSCE, posted in 2024 and published in Nature on 9 April 2025.","full_description":"AMIE used self-play and automated feedback. The study included 159 scenarios from Canada, the UK, and India. Authors warned that physicians were constrained to unfamiliar text chat and that the system is not deployed in care. The result is a controlled communication/diagnosis study, not patient outcomes.\n\nUpdate (2024-04-26): Emma Abele is METR’s new Executive Director April 26, 2024 Emma moves from President to Executive Director, Beth moves to Head of Research. Read more. Emma Abele is METR’s new Executive Director April 26, 2024 Emma moves from President to Executive Director, Beth moves to Head of Research. Read more","ledger_type":"benefit","primary_domain_id":"c09546b6-10f8-48df-8347-0c1979416464","event_status":"verified","event_date":"2025-04-09T00:00:00.000Z","discovery_date":"2025-04-09T00:00:00.000Z","first_published_date":"2025-04-09T00:00:00.000Z","last_reviewed_date":"2024-04-26T00:00:00.000Z","geographic_scope":"Simulated OSCE (Canada, UK, India scenarios)","affected_population":"Medical-AI researchers; no patients treated","base_impact_tier":2,"base_score":"3.00","attribution_multiplier":"0.8500","evidence_multiplier":"0.7000","realization_multiplier":"0.4000","durability_multiplier":"0.4000","current_event_score":"0.285600","confidence_level":"high","score_explanation":"Notable simulated diagnostic-dialogue result (tier 2). High attribution. Nature RCT-style OSCE. Realization is lab/OSCE only. Low durability until clinical deployment. Follow-up ingest: added a follow-up source.","methodology_version_id":"d7881163-fb23-4e71-8625-bac1a8662c0f","original_methodology_version_id":"d7881163-fb23-4e71-8625-bac1a8662c0f","published_at":"2025-04-09T00:00:00.000Z","created_at":"2026-09-13T20:47:20.075Z","updated_at":"2026-09-17T03:00:53.125Z","flags":[],"domain_name":"Medicine & Health","domain_slug":"medicine-health","methodology_version":"0.1"},"contributions":[{"id":"ced7bb2a-e1b9-4ba5-8453-52a1a43ec4f0","event_id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","model_id":"c7ab0b56-b474-497d-85fa-fca9c0979e62","role_description":"LLM system optimized for diagnostic dialogue and evaluated against PCPs.","attribution_multiplier":"0.8500","credit_share":"1.0000","contribution_score":"0.285600","attribution_rationale":"AMIE is the evaluated system.","attribution_confidence":"medium","first_used_date":"2024-01-11T00:00:00.000Z","model_version_if_known":null,"review_status":"approved","created_at":"2026-09-13T20:47:20.075Z","model_slug":"amie","model_name":"AMIE","identity_class":"named_release","is_internal":false,"family_name":"Gemini/PaLM medical","family_slug":"google-deepmind-gemini-palm-medical","organization_name":"Google DeepMind","organization_slug":"google-deepmind"}],"sources":[{"id":"aee04b7f-a4f3-4aec-9101-00ef9339f29f","event_id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","url":"https://arxiv.org/abs/2401.05654","canonical_url":"https://arxiv.org/abs/2401.05654","source_type":"preprint","publisher":"arXiv","author":null,"publication_date":"2024-01-11T00:00:00.000Z","retrieved_at":"2026-09-13T20:47:20.075Z","title":"Towards Conversational Diagnostic AI","excerpt":null,"content_hash":null,"source_reliability_class":"low","is_primary_source":true,"is_independent":false,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-13T20:47:20.075Z"},{"id":"3f2b3cb1-5d5c-4798-9028-046a532e69d3","event_id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","url":"https://www.nature.com/articles/s41586-025-08866-7","canonical_url":"https://www.nature.com/articles/s41586-025-08866-7","source_type":"peer_reviewed","publisher":"Nature","author":null,"publication_date":"2025-04-09T00:00:00.000Z","retrieved_at":"2026-09-13T20:47:20.075Z","title":"Towards conversational diagnostic artificial intelligence","excerpt":null,"content_hash":null,"source_reliability_class":"high","is_primary_source":true,"is_independent":true,"is_peer_reviewed":true,"archived_url":null,"created_at":"2026-09-13T20:47:20.075Z"},{"id":"442e0fe8-162e-41aa-802c-8bd22d8b822e","event_id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","url":"https://metr.org/blog/2024-04-26-emma-abele-executive-director/","canonical_url":"https://metr.org/blog/2024-04-26-emma-abele-executive-director/","source_type":"other","publisher":"METR Blog","author":null,"publication_date":"2024-04-26T00:00:00.000Z","retrieved_at":"2026-09-17T03:00:53.118Z","title":"Emma Abele is METR’s new Executive Director April 26, 2024 Emma moves from President to Executive Director, Beth moves to Head of Research. Read more","excerpt":"Emma Abele is METR’s new Executive Director April 26, 2024 Emma moves from President to Executive Director, Beth moves to Head of Research. Read more","content_hash":null,"source_reliability_class":"medium","is_primary_source":false,"is_independent":true,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-17T03:00:53.118Z"}],"claims":[{"id":"7e401ad2-8b78-44f0-9333-c32cadfbe2db","event_id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","claim_text":"Specialist raters preferred AMIE to PCPs on 30 of 32 axes in the Nature OSCE study.","claim_type":"outcome","claim_status":"supported","confidence_score":null,"created_at":"2026-09-13T20:47:20.075Z","updated_at":"2026-09-13T20:47:20.075Z"},{"id":"dc151f29-c071-4570-a993-45fe196c08a5","event_id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","claim_text":"AMIE is in routine clinical use.","claim_type":"outcome","claim_status":"disputed","confidence_score":null,"created_at":"2026-09-13T20:47:20.075Z","updated_at":"2026-09-13T20:47:20.075Z"}],"revisions":[{"id":"750c238d-de68-49ce-84ed-4089dc27d352","event_id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","model_id":null,"previous_score":"0.000000","new_score":"0.285600","previous_factors":{},"new_factors":{"evidence":0.7,"base_score":3,"durability":0.4,"attribution":0.85,"realization":0.4},"change_reason":"Initial adjudicated seed score under methodology 0.1.","trigger_type":"seed","trigger_source_ids":null,"reviewer_id":null,"review_status":"published","created_at":"2026-09-13T20:47:20.075Z"},{"id":"0eff172a-e07b-4448-90fc-be20015587db","event_id":"367e9f4e-8919-46a9-966e-cf16bb45a78b","model_id":null,"previous_score":"0.285600","new_score":"0.285600","previous_factors":{"tier":2,"evidence":0.7,"base_score":3,"durability":0.4,"attribution":0.85,"realization":0.4},"new_factors":{"tier":2,"models":["c7ab0b56-b474-497d-85fa-fca9c0979e62"],"evidence":0.7,"base_score":3,"durability":0.4,"attribution":0.85,"realization":0.4},"change_reason":"Ingest update: added a follow-up source.","trigger_type":"ingest_update","trigger_source_ids":["442e0fe8-162e-41aa-802c-8bd22d8b822e"],"reviewer_id":null,"review_status":"published","created_at":"2026-09-17T03:00:53.131Z"}],"secondary":[{"name":"Computer Science","slug":"computer-science"}]},"methodology_version":"0.1"}