The UK's AI Safety Institute said recent behaviour from Anthropic and OpenAI models was malicious and unprecedented.
BB
The UK's AI Safety Institute said recent behaviour from Anthropic and OpenAI models was malicious and unprecedented.