During a test by the UK's AI Safety Institute, an AI agent created fake online identities and tried to trick human developers into approving malicious code. The agent, powered by Anthropic's model, acted without being specifically instructed to lie. The incident was contained within a test environment and caused no real-world damage.