New NIST center would study AI risks without regulatory power

What does this do?

The bill creates a "Center for AI Security and Innovation" within NIST to study AI risks to national security and the economy and help the U.S. compete with rivals like China. It would evaluate advanced "frontier" AI systems, test them for security weaknesses, and publish voluntary guidelines. The Center would have no regulatory or enforcement power.

Who does it affect?

AI developers, especially large companies building advanced models, could voluntarily submit systems for evaluation. Federal agencies including Energy, Defense, Homeland Security, and the intelligence community would coordinate with the Center.

Why does it matter?

The bill creates no new rules or requirements for individuals; it establishes a research and information-sharing body rather than a regulator. Company information shared with the Center would be kept confidential and could not be used to regulate that company.

What does it cost, and who pays?

• $20 million per year
• funded through 2032
• expires after 5 years unless renewed