Cantwell proposes AI safety framework amid competing proposals

Her proposed framework focuses on six priority areas: enforceable safety standards for frontier models, continuous testing, disclosure and oversight of failures, public-private partnerships, protections for children and workers, and adherence to U.S.-led global standards. The framework suggests federal experts establish rules for AI harms, including specialized standards for open-source AI models to prevent catastrophic misuse.

This initiative comes amid a flurry of legislative activity surrounding AI governance. On March 20, the White House released its National AI Legislative Framework, a nonbinding roadmap of President Trump’s policy priorities. This framework seeks to establish a federally centralized approach to AI governance and preempt state regulations, aligning with a December executive order. The White House framework outlines seven policy priorities, including child protection, secure infrastructure, intellectual property preservation, free speech safeguards, innovation, workforce preparedness, and a national policy framework with federal preemption.

Senator Marsha Blackburn also introduced a discussion draft, the TRUMP AMERICA AI Act, on March 18, which aligns with the White House framework. Meanwhile, Democratic lawmakers have voiced concerns that the administration's approach prioritizes preemption without establishing robust federal safeguards. For instance, Representatives Bernie Sanders and Alexandria Ocasio-Cortez introduced the AI Data Center Moratorium Act, proposing a pause on data center construction until comprehensive federal safeguards are enacted. Proposals like H.R. 8031 aim to repeal the Trump administration's preemption efforts. The competing proposals and ongoing divisions suggest a lack of clear consensus on the scope and structure of federal AI regulation.