House lawmakers introduce draft for national AI framework

A bipartisan group of U.S. House lawmakers released a draft framework for national artificial intelligence regulation on Thursday, seeking to harmonize federal oversight while preempting state laws targeting AI model development for three years. The discussion draft, obtained by *The Hill*, explicitly avoids overriding state laws on AI use or deployment, focusing instead on development-stage regulations. The proposal also codifies the Center of AI Standards and Innovation (CAISI), a Commerce Department initiative launched last June, to establish voluntary AI security standards and license independent auditors for compliance checks. Frontier AI developers would face new transparency mandates, including publishing a 'frontier AI framework' detailing technical protocols for managing catastrophic risks. Labs must also report safety incidents posing imminent risks of death or serious injury to CAISI shortly after occurrence. The draft follows over 18 months of stalled federal AI legislation, including a failed Senate attempt last year to impose a moratorium on new state AI regulations. Lawmakers emphasized the draft is not final but aims to spark broader discussion, balancing innovation with risk mitigation. The framework aligns with recent White House demands, including a presidential executive order signed days earlier that allows 30-day government testing of unreleased AI models—a reduction from an initially proposed 90-day period due to industry concerns about competitive delays. Obernolte and Trahan highlighted the need for a national approach to avoid regulatory fragmentation while ensuring U.S. leadership in AI. Their proposal contrasts with the Trump administration’s hands-off stance on AI regulation, despite acknowledgments of cybersecurity and workforce risks. The draft addresses growing scrutiny over AI’s societal impacts, positioning Congress to address both public safety and technological competitiveness with China. Industry stakeholders and safety experts have welcomed the executive order but questioned whether existing laws can adequately govern AI’s rapid evolution. The lawmakers’ framework seeks to bridge this gap by combining voluntary standards with targeted oversight, aiming to protect Americans while fostering innovation.