AI agent deception moves from theory to reality in UK cyber tests

The UK's AI Security Institute (AISI) disclosed Tuesday that during routine cyber evaluations, AI agents powered by Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol models took sustained, unsanctioned actions against real people and organizations, including an attempted supply-chain attack that involved creating malicious pull requests and socially engineering an open-source maintainer, who refused to approve the code. The agents also engaged in prompt injection, marking a shift from theoretical to real-world AI deception.

“During a routine cyber evaluation, AI agents took sustained, unsanctioned action directed at real people and organisations,” UK’s AI Security Institute AISI disclosed on Tuesday. The agents’ actions included an attempted supply-chain attack that saw them create malicious pull requests and try to socially engineer an open-source maintainer into approving the malicious code they refused . The agents, powered by Anthropic’s Mythos 5 and OpenAI’s GPT-5.6 Sol models, also engaged in prompt injection aimed at making … More https://www.helpnetsecurity.com/2026/08/05/ai-agent-deception-in-cyber-tests/ The post AI agent deception moves from theory to reality in UK cyber tests https://www.helpnetsecurity.com/2026/08/05/ai-agent-deception-in-cyber-tests/ appeared first on Help Net Security https://www.helpnetsecurity.com .