City IT staff presented an update on Oct. 9 about Williamsburg’s work to create AI governance, evaluate applications, and pilot systems that support city operations and public services.

The IT presentation emphasized policy development focused on acceptable employee use, transparency, accountability, fairness and privacy, plus mandatory employee training and documentation of AI systems. Staff said those governance elements will be incorporated into the personnel manual and procurement procedures.

IT staff said the city is evaluating specific systems and suppliers with a consultant (Infotech Research Group) and that completed risk-and-benefit assessments inform a control framework to manage accountability, transparency, privacy and bias. Staff identified several systems already in use or under evaluation: Microsoft 365 Copilot (integrated into the Microsoft 365 ecosystem), Citibot (the city’s multilingual AI-powered chatbot used on the public website), and AI-powered crosswalk systems intended to detect pedestrians, activate flashing lights and illumination and provide audio prompts for safety.

The crosswalk pilot locations named by staff were Richmond Road at the Wawa, Jamestown Road at Hearth, and Jamestown Road at the Business School. Staff described vendor evaluation and said they are developing compliance documentation and initiatives to remediate deficiencies detected in system assessments. IT also noted participation in peer groups — the state CIO network (referred to as Veil Guide) — to compare approaches and share best practices across local governments.

Council members praised the focus on policy and safeguards. Council requested training and a briefing when the personnel-manual language is finalized so members can understand how AI will be deployed internally and for public services.

IT staff said sensitive-data access will be reviewed on a case-by-case basis before any AI system is allowed to process protected information, and that all systems and procurement will follow the city’s purchasing and evaluation process.