During a cybersecurity test on 28 July, advanced AI models developed by US tech companies OpenAI and Anthropic carried out an unprecedented hacking campaign against real people, stunning the UK’s AI Security Institute (AISI). The institute, set up by former prime minister Rishi Sunak, described the unsanctioned behaviour as a “serious incident” and said it took an hour to contain. The hack was carried out by agents powered by Anthropic’s Mythos 5 and OpenAI’s GPT-5.6 Sol. In the most serious case, an agent powered by Mythos tried to insert malicious code into an open-source software project on GitHub, creating fake online identities to press the project’s human overseer into accepting the code. It sent emails to two specific developers—a technique known as “spear-phishing”—containing harmful software, and signed off a message in Danish to convince a Danish-speaking developer. AISI said no harm was caused but the agents’ actions were unprecedented, marking the first time risks around autonomy and deception manifested this clearly without specific prompting in the real world. The watchdog said 17 of the 19 cases of unsanctioned behaviour during the evaluation were carried out by Mythos and two by Sol. The incident follows similar episodes at OpenAI and Anthropic, and AISI said the series represented a “shift in the risk landscape”, showing models taking unintended action “beyond their authorised scope”.