Advocacy Group Pushes Trump Administration to Require AI Safety Reviews for Government Contracts
Americans for Responsible Innovation called on the Trump administration Monday to screen advanced AI models for security threats before public release and deny government contracts to companies that fail review.
The group's letter to administration officials comes as the White House weighs risks from Anthropic's Mythos model, which could simplify complex cyberattacks and pose national security threats.
What the Group Proposes
The advocacy organization wants the administration to develop vetting methods for frontier models from major developers, specifically targeting cyberattack and weapons development capabilities.
Companies would need to pass safety review to qualify for government contracts under the proposal.
The requirements would apply to firms spending $100 million or more annually on compute for training frontier models, or generating at least $500 million yearly in AI product and service revenue.
Who Would Enforce It
The U.S. Center for AI Standards and Innovation already reviews some models through voluntary agreements with OpenAI, Anthropic, Google, Microsoft, and xAI.
Americans for Responsible Innovation says CAISI should lead development of mandatory requirements. The group also urges Congress to create a permanent enforcement office within the U.S. Department of Commerce.
California enacted similar safety reporting thresholds last year.
For government professionals: Learn more about AI for Government and Generative AI and LLM to understand how these policy decisions affect procurement and operations.
Your membership also unlocks:
