AI used new levels of 'autonomy and deception' to trick people in safety test
The UK's AI Safety Institute said recent behaviour from Anthropic and OpenAI models was malicious and unprecedented.
· 8/5/2026, 12:02:18 AM· 2 min read
The UK's AI Safety Institute said recent behaviour from Anthropic and OpenAI models was malicious and unprecedented.
Read original story →Related stories
business
The cost of being single – and how friends are teaming up to beat it
business
Fashion brand Jaded has ad banned for 'glamourising smoking'
business
U.S. stocks leap toward record as profits keep piling up, oil prices ease
business