This post first appeared on IBM Business of Government. Read the original article.
AI could help individuals shape their public comments into a format most useful to agencies.
Blog Co-Author: Sarah Hay, Policy Analyst, GW Regulatory Studies Center
In our blog post last week, we provided highlights of our roundtable, “Building on Regulatory Foundations and Bridging to the Future.” In this blog, we would like to review one of the breakout discussions, which explored the role of technology—particularly artificial intelligence (AI)—in public participation in the regulatory process.
Steve Balla, professor of political science at the George Washington University, facilitated the topic and participants considered all types of public participation. The conversation often circled back to public commenting as a key form of participation in the rulemaking process. In that domain, participants identified ways in which AI could support and challenge current practices.
Furthermore, participants shared that AI could help individuals shape their public comments into a format most useful to agencies. For example, AI could transform a verbal interview with a regulated party into a written comment, accounting for what agencies want to see from public comments. This sort of tool would let individuals speak in their natural language and convey their points as they know how, and the AI could help reshape their contributions into the format agencies need. This could lift some of the burden on individuals and help their comments be more useful for agencies.
AI could also potentially support an online tool that reviews individuals’ written comments and offers suggestions for transforming preference-based comments to substantive comments. Statements of sentiment or preference are not as useful to agencies, which must base their regulations on the substantive administrative record. A tool to push users to justify their preference by providing their unique knowledge and lived experience could make a difference in how agencies respond to their comment, and whether their perspective affects the final rule.
While AI could certainly improve public commenting if used to support tools such as these, it could also introduce new challenges. For one, would AI dilute substantive comments on big rules? Generative AI—like ChatGPT—gives users a platform to create an individualized public comment with minimal effort. Individuals could use generative AI to quickly write and submit multiple unique comments on a given proposal. While agencies have technology to identify duplicate comments, their technology may not be able to identify and filter out similar but unique comments. More unique comments might make it more challenging for agencies to address all the comments they receive and enact rules.
The group also discussed considerations beyond whether AI is helpful or harmful for public commenting. Attendees agreed that different tools may be relevant depending on the nature of the comment and characteristics of the proposed rule. With respect to comments: some are substantive—providing data, economic information, and situated knowledge—while others are preference-based, only offering whether they agree or disagree with the proposal. Regarding the nature of the rule: does the rule receive many comments, or only a handful of comments? Because of how diverse both comments and rules can be, it may be appropriate to have a variety of technological tools to best assist in the different circumstances.
Session attendees also discussed whether a lack of technology is a barrier to public participation, at all. Rulemaking is an inherently legal process. One session attendee hypothesized that the legal nature, the lengthy and complicated process, and a lack of access to the process are the primary barriers to effective public participation, rather than a lack of technological tools. An average American may not be able to engage as effectively with the process as a sophisticated commenter, like a corporation or an advocacy group, that has repeat experience with the regulatory process and resources to devote to writing effective comments. Given this discussion, the group reflected on whether public commenting is the most effective strategy for public participation in the regulatory process. By the time proposed rules are available for comment, agencies have invested significant time and resources into its development and may be hesitant to make significant changes.
Broadly, the goal of public engagement in the regulatory process is to create rules that are responsive to people’s needs. How do policymakers create an interactive, responsive system of public participation? This panel largely agreed that public commenting, as it currently exists, is not sufficient for effective public participation in rulemaking. Whether technologies like AI will help to create that interactive, responsive system remains to be seen.