The Senate Commerce Committee voted 18-7 Thursday to advance landmark artificial intelligence legislation that would require companies developing the most powerful AI systems to conduct safety evaluations before deployment.

The bipartisan bill, sponsored by Senators from both parties, represents the most significant attempt by Congress to regulate the rapidly evolving AI industry.

What the Bill Requires

Under the proposed legislation, companies training AI models above certain computational thresholds would be required to submit safety reports to a new federal oversight body within the Department of Commerce.

The requirements include pre-deployment testing for potential misuse, bias audits, and transparency reports about training data sources.