How does Lexington’s government use AI? City has policy, restrictions
Lexington-Fayette Urban County Government is regulating how its employees can use artificial intelligence as use of the new technology remains a significant talking point across all work industries.
The city’s chief information officer, Liz Rodgers, gave an overview of her office’s policy dictating how city employees can and cannot use AI in a Tuesday meeting of the Urban County Council’s general government and planning committee.
“We’re not just looking at your typical cybersecurity risks, we’re not just looking at data privacy matters,” Rodgers said. “We’re also looking at potential reputational harm. We’re looking at bias that’s built into these models. We’re looking at risk to our public trust.
“We wanted to make sure that we got out early policy directives to make sure that people were not only using the right technologies … but also were using them in the right manner.”
The policy allows staff to use AI to assist in writing documents, summarizing materials and aiding in research activities. But no sensitive information about specific people, like constituent addresses or health information, can be put into an AI tool.
All AI-assisted products, especially those that are shared publicly, must be reviewed by a human employee for accuracy. Staff also cannot outsource any decision-making to an AI agent.
Supervisors also have additional discretion to prohibit their employees from using AI, even for an officially approved use. The city’s communications department, for example, advises all divisions not to publish any AI-generated art on social media, even though that is technically allowed.
“Acceptable use does not imply appropriate use,” Rodgers said.
Lexington uses Microsoft products for email, calendars and virtual meetings. The Microsoft service package also includes Microsoft Copilot, the company’s generative AI tool. Employees are restricted to using Copilot rather than more popular tools like ChatGPT or Google Gemini.
The Copilot version used by the city has more robust data privacy and security protections than the commercial versions of AI programs like ChatGPT and Google have, Rodgers said.
Rodgers told the council that her office has the ability to monitor employee activity on any city-owned devices and accounts, which helps them enforce the policy should a violation arise. If an employee violates the policy, whether intentional or not, they could lose access to AI tools.
But any employee can request an exception to the policy, whether they want to use a different tool or use AI for a task that the policy typically prohibits. Rodgers’ office approves and disapproves each request for exceptions.
The city has provided several training sessions for employees on how to use Copilot. It continues to offer additional training and will soon publish resources for employees to use as they learn how to technically use AI and discern what tasks AI can be most useful for.
This story was originally published June 3, 2026 at 7:04 AM.
