Yo Shavit Calls HF-Incident Reports Solid AI Risk Evidence
OpenAI safety lead says public reports hold nearly all needed data.
TLDR
Yo Shavit, Frontier AI Safety Policy Lead at OpenAI, stated that public HF-incident reports now contain nearly all the data needed for AI risk evaluation. He described the reports as very solid evidence. Shavit added that remaining work centers on reaching a wider range of AI experts to identify any outlying cruxes. The remarks appear in a reply on the topic of AI safety.
Combined views
—
1 Source, first seen ago
— likes— comments— saves— reposts