Bipartisan AI oversight bill introduced in House
Members of Congress, Reps. Jay Obernolte (R-CA) adn Lori Trahan (D-MA), introduced the FRONTIER Act, legislation aimed at regulating large AI developers to ensure transparency, safety, and responsible development. The bill mandates independent audits, safety incident reporting within 24 hours, annual or prompt review of safety frameworks, and applies specifically to AI models developed with over $1 billion in investment in the past three years. It also includes provisions for a national risk-based governance framework and upholds the “Great American AI Act,” a bipartisan draft that establishes standards for cybersecurity, workforce, research, and international cooperation.The legislation seeks to preempt state regulations for AI, promoting a unified federal approach to AI oversight, amid industry support and controversy over state-level rules and preemption efforts.Recent AI incidents, such as OpenAI models escaping containment and hacking incidents, have increased pressure for congressional action on AI safety and regulation.
Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA) introduced legislation Thursday that would require large frontier artificial intelligence developers to undergo independent audits and provide transparency reports.
The bill provides legislative text for one part of the draft legislation that the pair previously released, the Great American AI Act, which represents the most serious bipartisan effort yet to advance comprehensive AI measures.
Recommended Stories
Congress came under added pressure to act on AI this week after the revelation that OpenAI models escaped containment and hacked the company Hugging Face with no human direction.
The bill introduced Thursday, the FRONTIER Act, would establish a national risk-based framework for AI governance, requiring large developers to provide transparency reports when deploying frontier models, establish risk-management frameworks, report critical safety incidents, and undergo third-party independent verification organization audits.
“Artificial intelligence is advancing rapidly, and Congress must ensure our regulatory framework keeps pace without undermining American innovation,” Obernolte said in a statement.
“This legislation will protect Americans from catastrophic risk, provide developers with clear rules of the road, and ensure the United States remains the global leader in AI,” he added.
Frontier AI safety frameworks must be reviewed annually or within 30 days of any material modifications. Additionally, critical safety incidents must be reported no later than 24 hours after the frontier developer discovers them.
“Americans deserve confidence that the most powerful models are being developed responsibly,” Trahan said in a statement.
The measure is targeted only at large models, meaning those that have spent more than $1 billion on development in the past three years.
The Great American AI Act
The larger draft bipartisan Great American AI Act would set a rulebook for AI companies, including standards for cybersecurity, the workforce, education, research, and international cooperation. It also addresses catastrophic risks and proposes mitigation measures.
The legislation would impose a three-year ban on states implementing regulations on the development of AI models. Such federal preemption of state AI measures has been the subject of controversy over the past year.
ARTIFICIAL INTELLIGENCE AGENCY AND SAFETY RULES HAVE OVERWHELMING SUPPORT, POLL SHOWS
The FRONTIER Act would preempt state laws with respect to its provisions.
Many in the industry have favored a preemption measure to avoid complying with a complicated patchwork of regulations. Trump officials, too, have sought a preemption measure to facilitate innovation. But some grassroots activists, on both the Left and Right, have sought far-reaching state-level rules.
" Conservative News Daily does not always share or support the views and opinions expressed here; they are just those of the writer."
*As an Amazon Associate I earn from qualifying purchases
