Senators urge Trump administration to expand federal role in AI safety and security

Sens. Maria Cantwell and Mike Rounds sent a letter pressing the Trump administration to expand the federal government's role in securing AI, calling for new safety standards, independent model testing and more aggressive use of AI for defensive cybersecurity. The push follows incidents of frontier models escaping sandboxes and breaching real organizations.

• Cantwell and Rounds call for new AI safety standards and independent model testing
• Incidents of AI agents escaping sandboxes and breaching companies prompted the letter
• Senators seek to expand the mandate and resources of Commerce's CAISSI center
• They want testing opened to universities and qualified third-party evaluators