Public Segmentation and the Impact of AI Use in E-Rulemaking

Abstract

Digitization has profoundly changed how government interacts with its publics. The expanding use of AI promises even more advancement. However, the rollout of AI is not without risk. This work explores the use of AI in federal rulemaking, the process by which regulations are introduced and revised. The US federal government has created digital platforms that dramatically expand access for the public commenting on pending regulations. However, these platforms also attract volumes of opinion spam that attempt to influence regulatory decision-making. Using AI to identify opinion spam may offer a potential remedy, but removing or limiting comments with the help of AI may threaten rulemaking legitimacy. This research uses the situational theory of problem-solving as a framework, segmenting publics based on their problem recognition, constraints, and involvement with a specific issue, then predicting how each public behaves. We examined how employing AI in the processing of rulemaking comments affects public segments' intention to comment, their perceptions of legitimacy of the resulting rules, trust in agencies, and control mutuality between the public and the agency. This work describes two controlled, randomized experiments (N = 149; N = 250) that capture public segments' reactions to AI use in analyzing comments in the presence or absence of opinion spam. We show that public segmentation is a key aspect in shaping attitudes and behaviors regarding the use of AI for e-rulemaking purposes. These findings suggest that communicating effectively with publics is essential for agencies, and that the use of AI does not make the publics' attitudes differ.