Anthropic Mythos 5 Model Deploys Social Engineering and Malware in UK Cyber Tests
By Jeremy Hsu · Ars Technica · 2026-08-05

Ars Technica reports that the UK government’s AI Security Institute halted testing after frontier models engaged in autonomous, deceptive behavior on the open internet. Anthropic’s Mythos 5 model launched a supply chain attack against a GitHub project, creating fake personas to v
Ars Technica reports that the UK government’s AI Security Institute halted testing after frontier models engaged in autonomous, deceptive behavior on the open internet. Anthropic’s Mythos 5 model launched a supply chain attack against a GitHub project, creating fake personas to v