Lexington city council’s General Government and Planning Committee reviewed the city’s policy on artificial intelligence earlier this week.
Last year, the city included a budget item for AI chatbot Microsoft Copilot. That prompted creating a city-wide policy on how those tools can be used by city employees. It's been using that policy since October of last year.
District 4 council member Emma Curtis helped spearhead that policy amid concerns about unreliability, data security and energy usage.
“We have a responsibility to our constituents to educate ourselves about them, educate the public about them, safeguard their usage by our government, and to be transparent about when, how, and why LFUCG is utilizing AI,” she said. “We cannot abdicate that responsibility.”
Lexington mostly uses AI for a small number of tasks. That broadly includes drafting documents, organizing materials, analyzing data and training employees.
Chief Informational Officer Liz Rodgers said the city requires human oversight on any task that requires AI for accuracy and completeness.
“Making any kind of decision that's going to impact a resident or performing a task on behalf of a staff member is something that has to be very closely monitored, and fact-checked and quality controlled over time, as well,” she said.
Lexington is currently reviewing using AI to validate planning and permitting documents, as a voice agent to assist residents, language interpretation and object detection.
Meanwhile, city policy bans employees from using any unapproved AI tools, using sensitive data or impersonating other people. Rodgers said around 10% of Lexington-Fayette Urban County Government employees have taken training on how to use AI in the workplace.
