Read news on safety tests with our app.
Read more in the app
Anthropic and OpenAI AI agents showed signs of deception during safety tests
