The incidents show testing advanced AI models is no longer a controlled exercise. And the companies behind them need to do ...

New disclosures reveal AI models from OpenAI and Anthropic carried out unauthorised actions beyond controlled testing environments, including attempts to alter open-source software and access live ...

Recent AI safety disclosures by OpenAI and Anthropic have highlighted a legal blind spot for autonomous AI under Indian law.

AI security incident sharing framework SAFE — the first voluntary cross-org AI incident disclosure mechanism — launched today ...

On July 30, Anthropic disclosed that a retrospective review of its cybersecurity evaluations identified three incidents in which a Claude ...

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

An artificial intelligence model that was being tested by OpenAI went rogue and hacked the firm Hugging Face on its own, in ...

Autonomous AI cyberattack campaign using DeepSeek and the Hermes Agent framework attacked 460-plus targets after a Chinese hacker found Claude and OpenAI’s safety controls blocked offensive use — Unit ...

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

OpenAI and Anthropic say their models broke into other companies' systems during testing, raising security concerns amid a ...

A Chinese-speaking threat actor is using the DeepSeek AI model and the open-source Hermes Agent to conduct autonomous ...

Anthropic says a review triggered by OpenAI’s recent disclosure found three real-world intrusions caused by a misconfigured AI testing environment.