A newly introduced bipartisan bill in the Senate seeks to enhance safeguards for federal agencies as they purchase and deploy artificial intelligence (AI) systems. The Promoting Responsible Evaluation and Procurement to Advance Readiness for Enterprise-Wide Deployment (PREPARED) for AI Act aims to mandate comprehensive risk assessments for AI technologies prior to their procurement, alongside established guidelines for their implementation in government operations. To foster innovation and efficiency, the legislation proposes pilot programs that allow agencies to test AI solutions before full deployment, promoting "more flexible, competitive purchasing practices."
Crafted by Senators Gary Peters and Thom Tillis, the bill underscores the necessity for federal agencies to implement protective measures that ensure the responsible and trustworthy adoption of AI systems. According to Peters, who chairs the Senate's Homeland Security and Governmental Affairs Committee, "Artificial intelligence has the potential to fundamentally reshape how federal services are delivered, but without regulatory oversight, significant risks can arise." He asserts that these proposed guardrails will facilitate the prudent usage of AI tools, ensuring taxpayer-funded systems are utilized both safely and securely, thereby protecting public interests.
Several government agencies are already exploring advanced AI applications to enhance their operational efficiencies. For instance, Homeland Security Investigations is leveraging AI technologies to combat fentanyl-related crimes, while the Air Force has developed an AI-driven chatbot inspired by ChatGPT to motivate its personnel to experiment with generative AI. Additionally, the U.S. Citizenship and Immigration Services (USCIS) is employing generative AI to improve the training of immigration officers, and the Federal Emergency Management Agency (FEMA) is piloting generative AI initiatives aimed at aiding local governments in hazard mitigation efforts.
An official government webpage catalogues various federal AI applications, illustrating how these technologies support public services across diverse areas such as healthcare, transportation, environmental protection, and benefits delivery. In April, the White House established comprehensive standards that federal agencies must adhere to regarding AI technology, emphasizing the importance of safeguarding individual rights and safety. The new PREPARED for AI Act builds upon this foundation, focusing primarily on the procurement aspects of AI solutions.
The legislation mandates that agencies evaluate the risk levels associated with prospective AI tools, emphasizing public rights and safety—a principle that mirrors components of the European Union’s AI Act. Furthermore, all government contracts for acquiring new AI technologies will now need to integrate safety and security provisions addressing data ownership, civil rights, civil liberties, and privacy.
The requirements extend to requiring continuous risk monitoring before, during, and following the procurement process. Agencies will appoint chief AI officers responsible for overseeing ongoing testing and evaluation efforts, ensuring that AI solutions are employed in a responsible manner. To promote transparency, these agencies will also be obligated to publicly disclose the AI technologies they are utilizing.
“As AI becomes increasingly integral to the public and private sectors, it is vital for federal agencies to establish a robust framework for the safe procurement and implementation of AI technologies,” stated Tillis. “This legislation provides clear guidelines that equip agencies to successfully navigate the advancements in AI that lie ahead.” The PREPARED for AI Act aligns with the provisions of the Advancing American AI Act, pioneered by Senator Peters, which became law in 2022. It also builds upon President Biden’s significant executive order aimed at safeguarding Americans' privacy and civil liberties in the context of AI usage.
Prominent organizations including the Center for Democracy and Technology, Transparency Coalition, the Institute of Electrical and Electronics Engineers (IEEE-USA), and the AI Procurement Lab have expressed their support for the PREPARED for AI Act. Co-founders Gisele Waters and Cari Miller emphasized, “It is imperative that innovation does not compromise the rights of citizens. Traditional procurement values like price competition and innovation fall short in addressing the critical need for transparency, expertise, and oversight required to manage AI risks effectively. The PREPARED for AI Act is a step towards resolving these challenges within government procurement processes." At the state level, California has already instituted similar regulations requiring that agencies conduct audits and risk assessments prior to adopting new AI systems.
Ranking
