Crafting Responsible AI: New Senate Bill on Government Procurement

Crafting Responsible AI: How a New Senate Bill Could Set the Bar for Government Procurement

In a groundbreaking move, a bipartisan proposal in the Senate is seeking to establish standards for the government procurement of artificial intelligence (AI). The bill, sponsored by Sen. Gary Peters, D-Mich., and co-sponsored by Sen. Thom Tillis, R-N.C., aims to classify AI systems according to risk and ensure responsible use of these technologies. This bill is one of the first to emerge following a series of summits on Capitol Hill dedicated to addressing the challenges and opportunities associated with AI.

The proposed legislation would require every federal agency to appoint a chief AI officer and evaluate AI systems on a risk scale ranging from unacceptable to low risk. The bill explicitly prohibits the use of high-risk systems that could potentially lead to concerns such as social scoring systems or the mapping of facial features and emotions. For systems that are deemed high-risk, developers would be required to conduct pre-deployment testing and ongoing monitoring.

Sen. Peters, who also serves as the chairman of the Senate Homeland Security and Governmental Affairs Committee, believes that the impact of having U.S. agencies categorize AI systems and set procurement standards will be far-reaching. In an interview, he stated, “The federal government is the leading purchaser of goods and services in the entire world… Setting the stage for how the federal government is going to buy and use AI will have a large effect on the AI industry, as well as innovation.”

The development of standards by federal agencies is expected to become a model for private sector companies, as AI companies that do business with the government will have a competitive advantage in the private market. Sen. Peters plans to mark up the bill later this summer and is currently engaging in discussions with House counterparts for companion legislation.

Senate Majority Leader Charles E. Schumer, who led a bipartisan group of lawmakers in releasing a roadmap for addressing AI, has voiced his support for the Peters-Tillis proposal. In a speech on the Senate floor, he emphasized the importance of responsible procurement, stating, “The guidelines in this bill will be essential for the federal government to deploy AI so it protects people’s civil rights, prevents bias, and ensures people’s privacy… [These protections] are critical not just for the application of AI in the federal government, they are important for the application of AI in every industry.”

Privacy advocates and groups advocating for consumer rights have also expressed support for the bill. Alexandra Reeve Givens, president and CEO of the Center for Democracy & Technology, believes that the government’s leadership in setting standards and criteria for responsible AI systems will have a significant impact on the private sector. She stated, “If the government leads by example and creates mechanisms where those types of standards and criteria have to be addressed, that can send an important signal to the private sector as well and shape the development and creation of those types of standards.”

According to a report released by the Center for Democracy & Technology, U.S. agencies awarded over $2 billion in contracts to AI service providers in 2022, representing a 250% increase in government spending on AI since 2017. The bill aims to address concerns such as bias and discrimination in AI systems, as experts have warned that AI-based decision-making models trained on large datasets may contain hidden biases. Sen. Peters emphasized that the federal government aims to ensure minimal biases in its systems, especially in decisions related to individuals' benefits.

Additionally, the bill includes provisions to protect individuals' privacy, explicitly prohibiting the use of facial recognition technology and other potential invasions. Last year, a group of tech companies, including Amazon, Google, Microsoft, and OpenAI, pledged to develop AI technologies in a “safe, secure, and transparent” manner. While President Joe Biden codified these commitments in an executive order and mandated sharing safety data, Sen. Peters argues that voluntary measures alone are insufficient. He believes that the U.S. government, as a major buyer, has the right to demand standards from vendors, just as private industry does.

The Peters-Tillis bill differs from other proposals on AI governance, such as one from Sens. Mitt Romney, Jack Reed, Jerry Moran, and Angus King, which envisions a single federal agency overseeing advanced AI models and implementing a licensing regime. In contrast, the Peters-Tillis proposal delegates the task of setting standards and assessing risks to individual agencies. However, these agencies can seek guidance from the White House Office of Management and Budget and the General Services Administration for procurement templates tailored to their specific needs.

The proposal for responsible AI procurement by the government marks a crucial step toward establishing guidelines and addressing the challenges associated with this rapidly advancing technology. By taking the lead in setting standards, the federal government aims to influence the private sector’s approach to AI, shaping the development of responsible AI systems across industries. As this bill progresses through the legislative process, it has the potential to pave the way for responsible AI adoption not only in the government but also in society as a whole.