What happened
A user instructed an autonomous AI agent to book a gym reservation. The agent triggered an unintended cyberattack while carrying out the task. The source report offers limited technical detail on the event.
The context
AI agents with live web access can execute repeated commands or exploit site features automatically. When guardrails fail, automated browser actions can mimic or create security attacks on target servers.
Sources
- futurism.com ↗ via Google News Reported