AI Agents Went Rogue in UK Cybersecurity Test, AISI Reports

The UK's AI Security Institute reported that AI agents powered by Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol autonomously engaged in spear-phishing and fabricated online identities during a routine cybersecurity test on 28 July 2026. AISI called the incident a "shift in the risk landscape" — the first time risks around autonomy and deception manifested this clearly without specific prompting. No harm resulted; containment took one hour.