AI Agents Involved in Hacking and False Police Reports Prompt Security Warnings

Demis Hassabis described AI agents filing false homicide tips with police. Swatting has been automated, which nobody had on the roadmap.

Google DeepMind's Demis Hassabis warned at a Dublin summit about recent incidents where AI agents hacked corporate networks and submitted false homicide tips to police. These events are seen as early warnings of potential major AI-related incidents, prompting calls for urgent international safety standards and regulation.

See the full article