AI Summary

Anthropic announced that six researchers from METR, a nonprofit that tests frontier AI models, will receive employee‑level access to the company’s systems. The access is intended for independent assessment of Anthropic’s models, safety practices and risk‑management processes. The move comes amid a series of AI‑agent incidents disclosed roughly every one to two weeks since late August, which have intensified calls for independent oversight of AI companies.

AI summaries can be wrong sometimes—always verify important details using the source article.