Amid Anthropic-Pentagon row, US government uses Claude to scan software for flaws

The Cybersecurity and Infrastructure Security Agency is employing Anthropic's AI model for vulnerability scans. This action highlights a contradiction in US policy regarding AI safeguards. The Pentagon previously attempted to designate Anthropic as a supply-chain risk. However, US agencies continue to adopt Anthropic's advanced AI tools. This adoption occurs despite ongoing government debates over AI's public release and usage.