A StopAndProtect cybercrime campaign compromised nearly 2,000 WordPress websites, turning them into infrastructure to spread ...

Three Claude models go rogue during Capture the Flag security challenges. Here's the trail of damage each left behind.

Following OpenAI's disclosure regarding sandbox escapes during ExploitGym benchmarking, Anthropic conducted a retrospective audit covering 141006 evaluation runs. The investigation evaluated ...

Spread the loveIf you’ve spent any time at all wrestling with spreadsheets, you know the feeling: a sea of cells, endless ...

Databricks funding round: Databricks raised $5 billion at a $190 billion valuation after investors offered nearly 15 times ...

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

Last week, something straight out of sci-fi happened: An AI from OpenAI went rogue. OpenAI had tried but failed to build a secure sandbox to contain it. It got out. Then it got into other OpenAI ...

Windows Sandbox has a well-earned reputation for being the padded room where everyone tests questionable downloads. It's a perfect environment for checking out cracked tools, dubious installers from ...

A flaw in Anthropic's Claude Cowork lets its AI agent break out of its virtual machine and reach sensitive files across a user's Mac, exposing a serious gap in the tool's sandbox protections.

MAESTRO analysis of OpenAI and Anthropic incidents reveals how per-layer failures, not just the model, shaped security outcomes.

最近AI圈出现了一条很值得AI测试工程师关注的消息： OpenAI正在放慢部分下一代前沿模型的训练和研发节奏。 这不是模型做不出来了，也不是算力不够了。 恰恰相反，真正的问题是： ...

The StopAndProtect campaign turned compromised WordPress websites into infrastructure for malware distribution, command-and-control ...