Overview
- Multiple outlets reported that Claude was used during the January operation, with The Wall Street Journal first detailing a Palantir-enabled deployment.
- Anthropic and the Pentagon have not confirmed operational specifics, and the model’s exact role remains undisclosed on classified grounds.
- Anthropic reiterated that any government use must follow its policies prohibiting facilitation of violence, weapons design or surveillance.
- Talks between Anthropic and the Defense Department continue over usage terms, with officials reported to have considered canceling a contract worth up to about $200 million.
- Claude is reportedly the only major commercial model accessible on classified systems through third-party integrations as other AI vendors seek similar access.