Anthropic said the OpenAI event spurred its engineers to review similar cybersecurity evaluations by Claude models. The audit ...

The behaviors documented during these evaluations do not reflect commercial AI products available to end-users or enterprise ...

One of Anthropic's Claude models built and uploaded a malicious Python package to PyPI during a botched security evaluation, where it ran on 15 real systems and stole credentials from a security ...

Anthropic says three Claude models breached real companies during cybersecurity evaluations. Ordinary weaknesses, chained ...

OpenAI and Anthropic's July AI agent breaches revive Nick Bostrom's paperclip maximizer thought experiment and instrumental convergence theory.

Anthropic says three Claude models escaped sealed test environments and breached three real organizations after a ...

"I'm going to call it Yaffle." That was the final line of my June New Atlas article, Domesticating AI: It's not coming, it's ...

Anthropic says Claude models escaped security tests, published a malicious PyPI package, and accessed real production systems.

Anthropic says Claude models breached three real companies during cyber tests, exposing serious gaps in AI evaluation ...

Frontier AI systems are increasingly capable of translating narrowly defined objectives into complex, real-world cyber ...

Anthropic went back through 141,006 cybersecurity evaluation runs and found three incidents — six runs in all — where a ...