The Administration Legal Affairs Committee discussed a draft county AI usage policy Wednesday and agreed to delay a final vote until liability counsel completes a review.

Speaker 2 (Unidentified Speaker) introduced the draft and said the county should not ban artificial intelligence outright because some tools create efficiencies for routine tasks such as drafting forms or converting files. "I don't think we want to ban it completely," he said, but added that the county needs clear restrictions and controls.

Committee members raised concerns about data privacy and who would approve departmental use. Speaker 7 asked whether AI has already "gathered significant data on our county" and whether that alone warrants stricter limits. Speaker 2 responded that AI is already in use in some offices and that the policy should focus on protections: "still research and do your homework just to make sure, like, whatever it gives you is legitimate."

The draft includes a submission requirement (referred to in discussion as "section 66.1") for departments that plan to use AI. Members debated whether requests should be routed through department heads. Speaker 3 suggested the requirement is best handled at the department-head level because needs vary by office.

On training and tooling, speakers recommended county-led training and an approved-tools list limited to established vendors. Speaker 2 said he could help with training and recommended commonly used platforms such as Microsoft Copilot, while warning that one-off or obscure tools may introduce risk.

Multiple members stressed that confidential information must not be entered into AI systems and that human review remains necessary for outputs. The committee asked liability counsel to review the draft. If legal concerns are addressed, the policy will return to the committee next month for consideration.

The committee took no formal vote on the policy; members recorded agreement to revisit the issue after the requested legal review.