Romney, Reed, Moran, King, Hassan Introduce Legislation to Mitigate Extreme AI Risks

WASHINGTON—U.S. Senators Mitt Romney (R-UT), Jack Reed (D-RI), Jerry Moran (R-KS), Angus King (I-ME), and Maggie Hassan (D-NH) today introduced first of its kind legislation that deals exclusively with safeguarding U.S. national security against extreme risks—chemical, biological, radiological, nuclear, and cyber threats—posed by future developments in advanced AI models. The bipartisan Preserving American Dominance in AI Act, crafted based on the group’s previously-released framework, would establish the Artificial Intelligence Safety Review Office within the Department of Commerce. The Office would work with industry to defend frontier models against misuse by foreign adversaries looking to exploit the technology for nefarious purposes.

The Office’s mission would be to prioritize the national security implications of AI while ensuring the U.S. domestic AI industry remains dominant over our foreign adversaries. This legislation is limited to frontier models—the most advanced AI models that are still yet to be developed.

“AI has the potential to dramatically improve and transform our way of life, but it also comes with enormous risks to national security. We must safeguard this advanced technology from terrorists and adversaries looking to cause widespread harm,” Senator Romney said. “My colleagues and I have spent the last several months turning our framework into bipartisan legislation that would preserve America’s competitive edge in AI while also ensuring the most advanced models do not pose unchecked chemical, biological, radiological, nuclear, or cyber risks. It is my hope that our legislation will serve as a starting point for discussion on what actions Congress and the incoming administration should take on AI-enabled extreme risks—without hampering American innovation.”

“While AI offers tremendous benefits, we must recognize the risks and take steps to ensure that this tool is not exploited by adversaries to threaten American security and leadership. This bipartisan bill would establish an AI Safety Review Office within the Department of Commerce to ensure the U.S. keeps pace with technology and responsibly supports innovation, opportunity, and discovery,” said Senator Reed.

“The evolution of artificial intelligence is an opportunity for U.S. innovation, efficiency and strategic advantage,” said Senator Moran. “However, we must responsibly harness the power of AI and make certain we are mitigating extreme risks that would threaten our national security. After receiving feedback from across the tech industry, government and other stakeholders, my colleagues and I have created landmark legislation that will help mitigate national security risks while making certain innovators are able to continue developing AI to secure a competitive edge over our adversaries in this critical technology area.”

“In the ever-evolving global threat landscape, the United States has to stay one step ahead of new technologies to protect both our national security and interests at home and abroad—and that means moving carefully, warily, and thoughtfully into an Artificial Intelligence future,” said Senator King. “This bipartisan Preserving American Dominance in AI Act provides critical guidelines for federal oversight of AI technology so that it cannot be misused by bad actors looking to cause harm. We must ask important questions now to wisely navigate our next steps and decisions. Thanks to my colleagues for working together on a solution that promotes American development while safeguarding the public against biological, chemical, cyber, and nuclear threats.”

“As artificial intelligence advances rapidly, we must work on a bipartisan basis to proactively address the potential national security threats posed by this technology,” said Senator Hassan. “This bipartisan legislation establishes safeguards to make sure that the U.S. remains the world leader in responsible AI. We must continue to work to protect against potential risks to public safety, such as the misuse of this technology by terrorists or criminal organizations. By implementing balanced oversight of the most advanced AI models, we can maintain our technological leadership while protecting public safety and our national interests.”

Background:

Artificial intelligence (AI) has the potential to dramatically improve and transform our way of life, but also presents a spectrum of risks that could be harmful to the American public, some of which could have catastrophic effects. Extremely powerful frontier AI could be misused by foreign adversaries, terrorists, and less sophisticated bad actors to cause widespread harm and threaten U.S. national security. Experts from across the U.S. government, industry, and academia believe that advanced AI could one day enable or assist in the development of chemical, biological, radiological, nuclear, and cyber weapons.

The Preserving American Dominance in Al Act would establish the Artificial Intelligence Safety Review Office, housed within the Department of Commerce. The Office would be responsible for working with frontier Al companies, large datacenters, and infrastructure-as-a-service providers (IaaS) to: (a) prevent exploitation of these industries by adversaries; and (b) ensure pre-deployment evaluations similar to CFIUS for the most advanced frontier Al models that pose chemical, biological, radiological, nuclear, and cyber risks.

Specifically, the bill would:

• Establish the Artificial Intelligence Safety Review Office, led by a Senate-confirmed Under Secretary of Commerce for Artificial Intelligence Safety.
• Direct the Office to:
  • Develop a take-home test for frontier Al model developers to evaluate their models for extreme risks prior to deployment;
  • Collaborate with the Al Safety Institute to provide industry with best practices and technical assistance; and
  • Study future Al-related risks and report findings to Congress.
• Require the most advanced data centers that train the most advanced Al models to report the owner and location of the facility to the Office.
• Protect industry from foreign adversarial actors by tasking laaS providers to implement know-your-customer standards for transactions with foreign persons.
• Safeguard frontier Al models from cyber breaches and intellectual property theft by requiring developers to implement cybersecurity standards.
• Help frontier Al model developers identify red-teaming best practices to test for chemical, biological, radiological, nuclear, or cyber risks.
• Provide the Under Secretary a 90-day window to review a frontier Al model’s safeguards against chemical, biological, radiological, nuclear, and cyber risks.
• Ensure a presumption of approval to prevent unnecessary roadblocks to deployment.
• Establish penalties for developers that fail to comply with the Act.

Full text of the Preserving American Dominance in Al Act can be found here and a one-pager can be found here.