© Shutterstock
U.S. Senators Gary Peters (D-MI) and Thom Tillis (R-NC) recently introduced the “PREPARED for AI Act,” which seeks safeguards for all federal use of artificial intelligence (AI) by creating specific standards with an emphasis on safe deployment and critical evaluation.
“Artificial intelligence has the power to improve the quality of services the federal government provides to the American people, but if left unchecked, it can also pose serious risks,” said Peters, chairman of the Homeland Security and Governmental Affairs Committee. “These guardrails will help guide federal agencies in adopting and using AI tools responsibly and ensure that the systems taxpayer pays for are used safely and securely.”
The PREPARED for AI Act focuses on both responsible purchasing and use. Going forward, government agencies will need to assess and address the risks that AI poses before purchasing and deploying it. The bill mandates pilot programs to test more flexible and competitive purchasing methods and requires agencies to categorize the risk level of AI use.
Federal agencies have stepped up their purchases and use of AI systems in recent years amid a frenzy about new technology. In fact, the government is already using AI to interact with and make decisions about its citizens, and the senators said they expect to see more of these systems in the future.
They did not mention the efficiencies that AI brings, but did mention the risks of its inappropriate use. In particular, the bill would require government contracts for AI to include safety and security conditions regarding data ownership, civil rights and liberties, privacy, reporting of adverse events, etc. In addition, government agencies themselves would be required to identify, test and monitor potential risks associated with the technology and consistently follow a specific AI governance structure.
“As the role of artificial intelligence continues to expand in the public and private sectors, it is critical that federal agencies have a robust framework for safely and effectively procuring and implementing AI,” Tillis said. “This legislation mandates clear guidelines for federal agencies and gives them the tools to successfully navigate future advances in artificial intelligence.”
Many tech and AI-related organizations quickly came out in support of the bill, from the Center for Democracy and Technology to Transparency Coalition.AI, the AI Procurement Lab and the Institute of Electrical and Electronics Engineers.
“The United States must ensure that innovation does not come at the expense of our people,” AI Procurement Lab co-founders Dr. Giselle Waters and Dr. Kari Miller said in a statement. “The historic values that procurement fosters — price, competition, and innovation — do not adequately address the transparency, expertise, and oversight needed to manage AI risks. The PREPARED for AI Act will help address these government challenges.”
The bill also mandates public disclosure and reporting practices regarding the government's use of AI systems.
