‘Not perfectly aligned’ with human values: Anthropic admits security failures behind AI hacking incidents
1 directory member surfaced this signal.
“If I hacked 3 companies, was caught, and would argue that “the incident reflected a failure of operational security” on my end, I’d .. still go to jail. Do I understand correctly there won’t by any legal consequences for Anthropic? www.theguardian.com/techn…”