US Senate Passes Bipartisan AI Safety and Innovation Act
The US Senate has passed the AI Safety and Innovation Act with strong bipartisan support, establishing federal standards for AI testing, transparency requirements, and a new regulatory body to oversee frontier AI development.
The United States Senate voted 78-19 to pass the AI Safety and Innovation Act, the most comprehensive federal AI legislation to date. The bill establishes mandatory safety testing for frontier AI models, creates transparency requirements for AI-generated content, and funds a new Federal AI Safety Board.
The legislation represents a compromise between lawmakers who prioritized safety regulations and those who emphasized preserving American competitiveness in AI development.
Key Provisions
The bill requires developers of frontier AI models exceeding defined compute thresholds to conduct pre-deployment safety evaluations and share results with the Federal AI Safety Board. It mandates watermarking of AI-generated images, audio, and video, and establishes liability frameworks for AI systems used in critical infrastructure.
A $10 billion research fund will support AI safety research at universities and national laboratories over the next five years.
Industry Reaction
Major AI companies including OpenAI, Google, and Anthropic expressed support for the legislation, calling it a balanced approach to regulation. Smaller AI startups raised concerns about compliance costs, though the bill includes exemptions for companies below certain revenue thresholds.
“This bill proves that safety and innovation aren’t opposites,” said the bill’s co-sponsor. “We can lead the world in AI while ensuring these systems serve the public interest.”
The House is expected to vote on the legislation within six weeks, with the White House signaling strong support.
]]>
