Every year, federal agencies publish thousands of new regulations. While agencies are obligated under various executive orders to conduct periodic reviews to identify regulations that might be outdated, redundant, or contain errors, that process can consume significant resources. Could artificial intelligence be of help, and what issues does using it raise?
Segal Family Professor of Regulatory Law and Policy Catherine Sharkey examined those questions in a report she prepared in May for the Administrative Conference of the United States (ACUS), an independent executive branch agency charged with issuing nonbinding recommendations to improve administrative and regulatory processes. Drawing on Sharkey’s report, on July 3, the ACUS published a recommendation in the Federal Register titled “Using Algorithmic Tools in Retrospective Review of Agency Rules.”
As described by the ACUS, the recommendation sets forth best practices for agencies to consider in their design and use of artificial intelligence or other algorithmic tools “to identify rules that are outdated or redundant, contain typographical errors or inaccurate cross-references, or might benefit from elaboration or clarification.” Additionally, it discusses how agencies can design such tools to promote “transparency, public participation, and accountability.”
A leading authority on administrative law, Sharkey is an ACUS senior fellow and a member of ACUS’s Roundtable on Artificial Intelligence in Federal Agencies. She served as a principal advisor for and co-author of a 2020 report issued jointly by the ACUS, NYU Law, and Stanford Law School titled “Government by Algorithm: Artificial Intelligence in Federal Administrative Agencies.” Sharkey was the sole author of her May report, “Algorithmic Tools in Retrospective Review of Agency Rules,” which was based on nearly 50 interviews she conducted with agency officials, industry stakeholders, and others.
Posted July 21, 2023