Houthis used Anthropic AI to attempt ballistic missile development; test fails

The Houthi group reportedly used Anthropic's AI to attempt building ballistic missiles, with a test apparently failing. This marks a notable first instance of a non-state actor leveraging frontier AI for weapons development, exposing gaps in AI safeguards. The failure limits immediate threat, but the precedent raises proliferation concerns.