Introduction to the Symposium on Artificial Intelligence and the Administrative Procedure Act
This post serves as the inaugural contribution to a symposium focusing on the intersection of Artificial Intelligence (AI) and the Administrative Procedure Act (APA). For additional posts in the series, click here.
The Growing Role of AI in Federal Agencies
In recent years, federal agencies have increasingly integrated AI and machine learning systems into their operations, with indications that the role of AI in administrative decision-making is expanding. The White House has urged for the acceleration of federal AI usage, prompting various agencies to adopt AI technologies.
For instance, the Department of Veterans Affairs leveraged AI to streamline contract evaluations, while the Secretary of Defense pledged to incorporate multiple AI systems within Pentagon networks. Notably, several agencies have granted their employees access to AI chatbots to enhance operational efficiency.
AI’s Influence on Policymaking
AI is now making significant inroads into policymaking. The Environmental Protection Agency (EPA) has announced plans to utilize AI for reviewing public comments on proposed regulations. Similarly, the Federal Energy Regulatory Commission (FERC) is encouraging AI applications to analyze grid policy. The Small Business Administration used a large language model to identify regulations impacting small businesses. Recently, the VA hinted at using AI in a final rule but did not disclose specific details.
In a more ambitious proposal, the U.S. Department of General Services suggested using AI to rescind half of all federal regulations, while the Department of Transportation is reportedly considering Google Gemini for drafting rules.
Questions Surrounding Government Use of AI
The array of questions regarding government AI usage is nearly limitless, touching on issues of data governance, privacy, bias, conflicts of interest, procurement, civil service, and public trust. This symposium specifically addresses administrative law, focusing on the implications of the APA, which governs the exercise of discretion and judgment by agencies in executing statutory directives from Congress.
One key question arises: What occurs when long-established good government doctrines of administrative law are applied to agency actions that rely on new algorithmic systems? It may be that agencies simply adopt the latest technology to enhance their capabilities while fulfilling administrative law requirements. Alternatively, AI tools could undermine state capacity, reduce political accountability, and lead to arbitrary policy decisions. A combination of both outcomes is also possible.
A Symposium for All Perspectives
This symposium caters to AI skeptics, believers, and everyone in between. A diverse series of essays will be presented, featuring insights from both academics and practitioners. Some contributions will address core doctrinal questions regarding how agency use of AI interacts with administrative law, particularly in light of rapidly advancing technology. Others will consider normative questions about how agency use of AI aligns with the objectives of administrative law and the responsible realization of its benefits.
Additionally, reports will be presented on how regulators in the United States and abroad are adapting to AI and deploying it prudently. The symposium aims to initiate discussions that could help clarify the pressing questions surrounding this evolving field.
