Story perspectives

AI System AMIE Matches Doctors, Outperforms on Complex Cases

6/25/2026

Summary

Story summary

A blinded virtual OSCE study in the United Kingdom compared AMIE with 21 primary-care physicians over 100 cases.
Specialist reviewers judged AMIE’s management reasoning non-inferior to the physicians’.
AMIE scored higher on treatment precision, investigation selection, and guideline alignment.
On the RxQA medication-reasoning benchmark, AMIE outperformed physicians on higher-difficulty questions.
AMIE runs on Gemini’s long-context model, using in-context retrieval and structured reasoning anchored to current guidelines and drug formularies.

1 of 1

Story summary

A blinded virtual OSCE study in the United Kingdom compared AMIE with 21 primary-care physicians over 100 cases.
Specialist reviewers judged AMIE’s management reasoning non-inferior to the physicians’.
AMIE scored higher on treatment precision, investigation selection, and guideline alignment.
On the RxQA medication-reasoning benchmark, AMIE outperformed physicians on higher-difficulty questions.
AMIE runs on Gemini’s long-context model, using in-context retrieval and structured reasoning anchored to current guidelines and drug formularies.