UK experts sound alarm after AI caught trying to trick human with malicious code

A powerful AI agent created fake online identities in an effort to trick a human into giving it access to a popular online development platform – and sabotage it with malicious code.