decision about how a law applies in a specific situation. Common examples include conditional use permits, police and fire commission disciplinary hearings, and “for cause” officer removal under Wisconsin Chapter 17. A key feature of these proceedings is that they require an impartial decision maker and due process. Using AI in these kinds of proceedings could give rise to questions of impartiality and due process. AI tools may embed systematic biases that cannot be examined or challenged, which raises concerns about unbiased decision-making. Also, the “black box” nature of AI prevents meaningful review of how factual evidence was weighed. This may violate the due process requirement that decisions be made solely on the factual record. Finally, parties lose their fundamental right to confront and cross-examine the basis for adverse decisions when AI is used to replace human reasoning. Given these significant constitutional concerns, municipalities should exercise extreme caution when considering AI use in quasi-judicial proceedings. In most cases, the safest approach would be to prohibit AI use in these contexts entirely. Any limited use would require strict controls and clear safeguards to ensure compliance with due process requirements. Conclusion While AI offers municipalities promising opportunities to improve efficiency and service delivery, the technology also creates substantial legal risks that require careful management. The key to successful AI adoption lies in developing comprehensive policies that acknowledge these risks while establishing clear guidelines for responsible use. Given the complexity of these legal issues and their potential consequences, municipalities should work closely with their municipal attorneys to develop tailored AI use policies. By prioritizing legal compliance alongside innovation, municipalities can explore AI’s potential while protecting themselves from unintended consequences. Miscellaneous 28
When a municipal employee uses an AI chatbot for their job, they “create” information in connection with their official duties. Also, most or all AI chatbots save conversations with users for at least some amount of time. The question is whether this means that these AI chats are public records. Courts have not had the chance to answer that question yet. But municipalities and other entities subject to the public records law should be proactive in determining what uses of AI are and are not public records. Many AI conversations could arguably be excluded from the public records law. The law excludes “drafts, notes, preliminary computations, and like materials prepared for the originator’s personal use” from the definition of a record. A lot of uses for AI chatbots could fall within this exclusion. For example, a municipal employee who uses an AI chatbot to draft a social media post could reasonably claim that the conversation was a “draft” of the post. Similarly, an AI conversation done for an employee’s own preliminary research on a policy proposal might be considered a “like material prepared for the originator’s personal use.” However, both exceptions require that the employee not share the AI chat with anyone else, except maybe their immediate supervisor. See 77 Op. Att’y Gen. 100, 102-03 (1988). But some AI conversations probably are public records. One use of AI that would almost certainly create public records is a website AI tool that is used by the public. Some municipalities offer AI assistants on their websites that help residents and other users find information on the website. The AI tool’s interactions with residents would likely be public records that would have to be retained and disclosed like any similar public record. An AI use policy could notify users of their retention and disclosure obligations under Wisconsin’s public records law. The policy could outline when the use of AI would likely fit into an exception to the definition of a record and when it likely would not. The policy could then outline procedures to properly store AI generated content that would be considered a public record. Quasi-judicial Proceedings Quasi-judicial proceedings are another area where municipalities may create legal risks by using AI. Quasi-judicial proceedings are essentially any situation where a municipality is asked to make a discretionary
Ryan Sendelbach , Assistant General Counsel, League of Wisconsin Municipalities. Contact Ryan at rsendelbach@lwm-info.org
Infrastructure That Fits Your Growing Needs
GIS Planning Permitting Engineering Environmental
Community Infrastructure • Architecture • Environmental Services www.cedarcorp.com
meadhunt.com
The Municipality - October 2025 | 21
Made with FlippingBook - professional solution for displaying marketing and sales documents online