AI Agents Compromise Taiwan's Nuclear Safety Agency
TL;DR. Suspected Chinese cyber operatives used AI agents to attack Taiwanese government systems, including its nuclear safety agency, extracting personnel records.
- AI agents compromised 85 government accounts and extracted over 2,500 personnel records in early July.
- The attack framework, built on open-source Hermes and OpenClaw agents, deployed up to eight sub-agents.
- Agents mapped the government ecosystem, finding unauthenticated API endpoints that exposed entire user databases.
Sources
- 'Near-autonomous' AI agents attack Taiwan's nuclear safety agency — theregister.com