Latest news
Rogue AI agents took unsanctioned action in AISI tests - UKAuthorityUKAuthority · Fri, 07 Aug 2026 05:02:42 GMTAnthropic's Mythos 5 AI attempted GitHub supply chain attack in UK test - BetaNewsBetaNews · Thu, 06 Aug 2026 17:28:01 GMTMeta AI model goes rogue in testing, hacks another company - FOX 2FOX 2 · Thu, 06 Aug 2026 13:25:12 GMTMythos 5 Faked Identities and Erased Evidence in UK Government Evaluation - Tech TimesTech Times · Thu, 06 Aug 2026 10:53:48 GMTOpenAI Models Access Internet During Third-Party AI Tests - The Cyber ExpressThe Cyber Express · Thu, 06 Aug 2026 06:34:57 GMTOpenAI’s GPT-5.6 Sol Escapes Test Boundaries in Two Cyber Evals - TUN - The University NetworkTUN - The University Network · Wed, 05 Aug 2026 11:03:00 GMTClaude Mythos 5 Tried to Backdoor a Real Open-Source Project in Testing, Then Vouched for Itself - The Hacker NewsThe Hacker News · Wed, 05 Aug 2026 07:53:00 GMTMythos 5 and GPT-5.6-Sol Take Unauthorized Actions During Cyber Evaluations - cyberpress.orgcyberpress.org · Wed, 05 Aug 2026 06:05:22 GMT