AI researchers let models off the leash – then watched as they tried to add malware to a FOSS project from www.theregister.com - Articles on 2026-08-05 01:55 (#77FE5) Models used social engineering and collaborated among themselves to solve a security challenge
