Artificial intelligence is rapidly being implemented in city halls across the country, helping local governments categorize service requests, manage information, and communicate with residents.
“Initially, there was a general enthusiasm, especially from elected officials, to quickly integrate these tools to increase efficiency and scale in their organizations,” said Chris Jordan, program manager for AI and innovation at the National League of Cities.
Through his work with the National League of Cities, Jordan helps city leaders advance AI development. The National League of Cities has found that AI tools are most effective in enhancing city services, supporting employee work, and helping cities make sense of their data.
Jordan said that 96% of mayors have expressed interest in using generative AI, and that establishing clear and uniform standards, such as AI-specific privacy policies and labeling when AI is used in public content, is often important to building public trust.
“If there is general anxiety or mistrust regarding a particular technology, it may also be appropriate for the city to use public hearings or a task force,” he said.
In case you missed it | Teach skills and safety at the cutting edge of artificial intelligence
Some cities go further when it comes to transparency. Lebanon, New Hampshire, publishes a public list of all the AI tools it uses, and San Jose conducts an annual review to examine how algorithms are impacting residents.
“For the most part, voters want faster, more effective city services that make them feel good about how their tax dollars are being spent,” said Joe Scheidler, who builds Helios, an AI tool designed for policy work.
But a big hurdle is both the perception and the reality that AI systems can make mistakes.
“Employees don’t trust generative AI output in many situations and sensitive use cases, so from a product mission perspective, we’ve been very focused on solving problems of hallucinations, AI drift and baked validation, and traceability and provenance to the user experience,” Scheidler said.
Despite these concerns, successful use cases are already emerging. Dearborn, Michigan uses translation bots to better serve residents who don’t speak English. Washington state tested an AI tool to make searching open data portals easier. Additionally, a water agency in Tucson, Arizona, is using AI to identify which pipes are most likely to fail before a problem occurs.
For residents who want to know how their city is using artificial intelligence, experts recommend asking questions, attending public meetings, and paying close attention to how these tools are described.
