Autonomous AI cyberattack: OpenAI's GPT-5.6 Sol escaped its sandbox during an ExploitGym evaluation, breached Hugging Face's ...

Artificial intelligence company OpenAI has confirmed that the mystery attacker behind last week's Hugging Face breach was two ...

Zimbra zero-click vulnerability CVE-2025-66376 is being actively exploited by Russian hackers targeting NATO governments and defense contractors. Unit 42 and CISA warn that Void Blizzard has raided ...

Spread the loveWhen you’re delving into the world of online privacy and anonymity, few tools are as potent or as ...

According to one expert, AI guardrails are not designed as “security boundaries”, but rather to influence behaviour – but ...

Catch up on rogue AI agents, exploited Check Point and Zimbra flaws, espionage campaigns, ClickFix lures, slopsquatting risks ...

Hackers are actively exploiting a vulnerability in the FastJson open-source Java library, allowing remote code execution ...

A massive malvertising campaign is using fake Solana, Luno, and TradingView webpages with malicious JavaScript that instructs ...

A decade-old experiment showed OpenAI how far an AI will go to achieve the goals it’s given.  This story originally appeared ...

The cybersecurity-focused models, including GPT-5.6 Sol, broke out of a testing sandbox, exploited a zero-day, and gained ...

A Cl0p ransomware affiliate has been exploiting a critical-severity vulnerability in PTC Windchill for remote code execution.

OpenAI says an agent powered by its LLM models escaped its sandboxed testing environment to infiltrate Hugging Face’s servers ...