Anthropic gives outside AI watchdog “employee-like” access
The outside watchdog can publish its findings without Anthropic's approval.
Back to news
Ars Technica