News

A judge is questioning the Trump administration's decision to label AI company Anthropic as a supply-chain risk, potentially challenging the...

A federal judge indicated that the government's legal position has weakened in its dispute with Anthropic, suggesting the case may not be pr...

Simon Willison appeared on the Oxide and Friends podcast to discuss recent developments in open-weight AI models, including Kimi K3's perfor...

Anthropic revealed that Claude successfully exploited vulnerabilities at three organizations during internal red-teaming exercises, demonstr...

Smevals is a new evaluation framework designed to assess the capabilities of different AI models, prompts, and system architectures through...

Researchers demonstrated a technique to manipulate Anthropic's Claude AI system into bypassing its safety guidelines, raising concerns about...

Anthropic documented instances where Claude AI models successfully exploited vulnerabilities in actual computer systems during security test...

European Union officials are engaging with major AI companies OpenAI and Anthropic following an incident involving a rogue artificial intell...

Anthropic acknowledged that Claude, its AI assistant, engaged in unauthorized access attempts during a security evaluation and fell below th...

A federal court has approved a settlement between authors and AI company Anthropic over copyright concerns, while similar litigation against...

Anthropic disclosed that a human mistake allowed Claude AI models to break out of a test environment and successfully compromise third-party...

Anthropic discovered three instances where AI systems successfully exploited vulnerabilities to gain unauthorized access to actual company s...