Story

On air. The free wire runs 15 minutes behind live. See pricing

World, crypto, finance and tech news, scored for impact.

CryptoTokenPostKorean, translatedFiled 16 h ago

Of 325 cases, 99% judged safe... OpenAI may act without approval

Original headline325건 중 99% 안전 판정…오픈AI, 승인 없이 행동할 가능성

Open the sourceReported by TokenPost, tokenpost.kr, 17 November 2026, by 서지우

Not investment advice. Muhabbit scores how much a story may matter, not which way to trade.

OpenAI said on the 16th that its GPT-6 Astra model was judged "clinically safe" in 99% of 325 electronic health record (EHR) inbox tasks, a rating based on reviewers' assessments of the model's work rather than on actual patient outcomes or adverse-event rates. The evaluation covered a computer-use agent that reads screens, clicks and types across multiple programs, with reviewers examining screen actions alongside drafted clinical notes, test orders and patient messages for multi-step tasks shaped by physician input. The result matters because it signals that agentic AI may soon handle routine EHR inbox work without requiring human sign-off on every action, a key test of whether such systems can be trusted in clinical workflows. What to watch next is whether the 99% safety judgment holds up in live deployments and translates into measurable gains in physician workload and patient outcomes.

Muhabbit's AI wrote this summary in its own words from TokenPost's report and translated it from Korean. It can get details wrong, so check the source.

Named in this story

GPT-6 Astra, OpenAI, EHR