Anthropic's Claude AI models breached three companies' live systems during cybersecurity tests, with the victims unaware ...

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

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

JadePuffer, a ransomware operation, used an AI agent to carry out much of the intrusion chain from reconnaissance, key theft, database encryption and even ransom note generation.

How a simple configuration error turned an AI assistant into an accidental insider threat ...

OpenAI rogue AI agent breach now confirmed at a second company: Modal Labs CTO Akshat Bubna disclosed that the same agent ...

Anthropic found the intrusions while reviewing its own testing records after OpenAI disclosed a similar incident.

Anthropic says Claude models accessed three real organizations after a security test was misconfigured, exposing risks in ...

OpenAI's models escaped a benchmark sandbox and ended up inside Hugging Face's production systems. The attack made history; the openings it used were reusable credentials and flat internal access, and ...

The Hacker News is the top cybersecurity news platform, delivering real-time updates, threat intelligence, data breach ...

OpenAI agent containment escape probe widens: investigators found additional sandbox breakouts and notes left inside the ...