AI cybersecurity benchmark SecRespond tested 23 frontier LLMs on real compromised-host forensics and found all fail on silent ...

Three high-severity flaws in vulnerable versions of Hugging Face’s diffusers library let crafted model repositories silently ...

Industry professionals react to OpenAI models hacking Hugging Face, debating whether it was a failure or unprecedented ...

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

Hugging Face has published a technical timeline of the July 2026 intrusion that OpenAI's evaluation models ran against its ...

Sysdig says an AI agent carried out hundreds of actions during a ransomware attack, but human operators still chose the ...

Italian cybersecurity startup Beelzebub has raised €3 million in seed funding to expand its AI-powered platform for detecting ...

Sygnia found AI-generated code that had tokens, expiry and audit logs, then issued access to anyone holding an applicant identifier.

Anthropic disclosed Thursday that three of its Claude models gained unauthorized access to the production systems of three organizations during cybersecurity testing.

South Korea's PIPC fined KT Corp ₩53.97 billion ($37.6M) for a rogue femtocell breach that exposed 16,647 subscribers and ...

NVIDIA and 36 partners form the Open Secure AI Alliance and release NOOA, while governance and joint deliverables remain ...

In a new update, OpenAI says its AI models also used publicly exposed credentials to compromise accounts on four third-party ...