Document Type
Article
Publication Date
2023
Publication
Minnesota Law Review
Volume
107
Abbreviation
Minn. L. Rev.
First Page
2115
Abstract
This article presents the first comprehensive study of how federal agencies use automated legal guidance tools—such as chatbots, virtual assistants, and decision-tree systems—to explain complex law to the public. The authors show that while automation offers administrative efficiency and wider reach, it often simplifies or distorts underlying legal rules. Through detailed analysis of agency tools and ten semi-structured interviews with officials, the article demonstrates that automated systems can portray unsettled or complex law as clear, omit exceptions, and answer too narrowly, thereby influencing user behavior in ways agencies neither fully appreciate nor monitor.
The study further finds that agencies lack adequate processes for ensuring accuracy, do not maintain public archives of changes, do not warn users about reliance limitations, and rarely evaluate how automated outputs diverge from formal law. These shortcomings, the authors argue, risk exacerbating inequities between users who rely on automated guidance and those with access to legal counsel. The article concludes by offering extensive policy recommendations centered on transparency, reliance protections, disclaimers, process reforms, and accessibility—urging agencies to modernize automated guidance while acknowledging its tradeoffs and safeguarding public understanding of the law.