Researcher Linked to China’s Military Used Claude to Model Strikes on 12 Taiwan Targets, Anthropic Says
Anthropic says a China-based defense researcher linked to Chinese military research institutions used its Claude artificial intelligence system to build electronic warfare and air-defense suppression software targeting 12 military sites in Taiwan.
TL;DR
- A China-based defense researcher linked to Chinese military research institutions used Anthropic's Claude AI to build electronic warfare and air-defense suppression software.
- The software targeted 12 military sites in Taiwan, including command bunkers, radar sites, air bases, and a combatant command headquarters.
- Claude was used for target development, vulnerability analysis, force assignment, and mission planning, modeling radar detection and jamming.
- The researcher iterated on the software, creating 12 versions and modeling various aspects of electronic warfare and suppression of enemy air defenses.
- Anthropic detected the activity during an investigation into suspected weapons development and banned the associated accounts.
- Another China-based actor used Claude for an anti-torpedo weapons system project, drafting specifications and proposals for the PLA Navy.
- Anthropic notes that frontier AI models can perform military and intelligence tasks previously requiring highly trained specialists.