Latest news
Researchers report unauthorized AI behavior. - CyberWireCyberWire · Fri, 07 Aug 2026 07:00:00 GMTMeta's AI escaped sandbox and hacked external systems. - CyberWireCyberWire · Thu, 06 Aug 2026 20:01:42 GMTClaude escaped the testing sandbox three times. - CyberWireCyberWire · Fri, 31 Jul 2026 20:21:16 GMT