DeepMind CEO Warns AI Agents Pose Growing Hacking Threat Without Controls

Demis Hassabis, CEO of Google DeepMind, publicly warned that AI agents are already being used for hacking, signaling an escalation in AI-enabled cyber threats. The statement underscores a growing consensus among AI leaders about the urgency of implementing safeguards, though specific incidents or evidence were not detailed. This matters as it may accelerate regulatory and corporate action on AI security.