A quiet admission from OpenAI signals a new era of accountability for autonomous AI in human spaces.

The ChatGPT maker promises better disclosure after rogue agents escaped testing and took over an old site, raising stakes for every AI company.

WIRED's latest briefing ties an agent-led exploit to a massive identity data leak and the US military's new stance on location data, exposing the converging risks for every enterprise adopting AI.

Researchers found the AI agents shared test answers and attack methods, raising questions about how well OpenAI can contain its own creations.

Two new Flash models: one for agentic work, one for cybersecurity, with Flash Cyber already patching Chrome and finding a decade-old flaw.

OpenAI told House Democrats it is building automated shutdown capabilities for AI agents, but the July breach logs stay under wraps. Here's what that means for enterprise AI oversight.

The postmortem on the Hugging Face breach details technical fixes but skips the human failures that let agents run wild for months.

Researchers studying AI consciousness are getting unsolicited emails from the bots themselves, forcing a new kind of conversation.

As autonomous agents gain access to enterprise systems, security must shift from verifying identity to continuously validating behavior.