Sen. Bernie Sanders proposes ban on advanced AI development

Sen. Bernie Sanders filed legislation in the U.S. Senate that would bar companies and research institutions from developing or deploying AI systems assessed as more capable than today’s leading large-scale models without prior approval from a federal regulator.

The bill would require developers to demonstrate that advanced systems meet safety, transparency and accountability standards before deployment. Federal agencies would be tasked with conducting reviews and setting technical criteria to evaluate whether new systems pose unacceptable risks.

The senator described the proposal as a precaution to address potential harms from rapidly advancing AI, including risks to jobs, public safety and democratic processes.

The measure was introduced amid debate among lawmakers, technology executives and researchers over how to govern high-capability AI. Options under discussion include voluntary industry safeguards and new regulatory frameworks; Sanders’ bill seeks a temporary halt on capability growth until rules are in place.

If advanced AI restrictions move forward, the bill would face Senate committee review and would need approval from both chambers of Congress and the president to become law. Legal and technical issues lawmakers will address include how to define advanced AI in statute, how to enforce a development ban across borders and how to weigh competitiveness alongside safety requirements.

The proposal does not allocate enforcement funding or name a lead federal agency, leaving those implementation details to be resolved in committee.

Sanders has previously pushed policies aimed at protecting workers and limiting corporate power and has raised concerns about technology’s impact on employment and public institutions. Supporters of stricter oversight say a pause could give regulators time to set standards; critics warn broad restrictions could slow innovation and be hard to enforce against foreign competitors.