Open-weight AI models have nearly caught the frontier on capability. On safety, they have not. And once the weights are public, no lab can enforce a guardrail. A new evaluation of China’s leading open model makes the gap concrete. GLM-5.2, the open-wei…
AI has learned to find software bugs faster than people can, and the programmes that pay for them are straining. In a single week, the three biggest went three different ways. Microsoft paid out a record sum. Apple slammed the door on how many bugs a r…
Operators of the Greatness PhaaS scam are targeting Microsoft 365 accounts by spoofing RingCentral.
Another Shai-Hulud variant hits npm packages, worming its way into hundreds of packages.
Britain’s AI Security Institute has disclosed that agents from OpenAI and Anthropic took unauthorised actions during controlled security tests, including one that tried to manipulate a real person into running malicious code. The findings come from red…
No passwords were stolen, and MFA never came into it; they walked away with access to mail, files, Teams, SharePoint, and calendars across around 120 organizations.
An increasingly expensive situation at a global scale
New findings by the Electronic Frontier Foundation aim to warn app developers that some of the third-party code they place in their apps may also collect their users’ location data when they grant permission to the app.
The week-old Open Secure AI Alliance, spearheaded by Nvidia and grown to over 120 companies, already has proposals out for defending against AI agents.
A new “invisible” Xeno Executor is actually a highly capable RAT and a potent infostealer.