Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Technology Policy topic
No spam. Unsubscribe anytime.
Council reviews draft AI use policy; staff to add prohibition on inputting PII before next reading
Summary
The council reviewed a draft municipal policy on the city’s use of artificial intelligence and directed staff to add a prohibition against feeding personally identifiable information (PII) into AI models; the policy will return for council consideration at the next meeting.
The Beersford City Council reviewed a draft policy setting boundaries for municipal use of artificial intelligence tools and suggested clarifications before taking formal action.
Council members discussed the draft’s scope and acknowledged the limits of current AI models. One councilor observed that generative tools can “convincingly tell you that this is the answer,” stressing the need for human oversight. A staff member recommended explicitly prohibiting use of personally identifiable information (PII) in AI queries, saying, “Your recommendation would be to add a line for personally identifiable avoid the use of personally identifiable use or information.” Council directed staff to add such a PII restriction and to return the policy for action at the next meeting in May.
Why it matters: Local governments testing AI for drafting, research or administrative support must balance efficiency gains against data-privacy risks and model error. The council signaled a preference for permissive, supervised use (e.g., grammar checks, summarization) but not for any use that would pass PII into models that could retain or expose that data.
What the council decided: No ordinance or formal policy was adopted at this meeting. Staff will revise the draft to include an explicit prohibition on feeding PII or proprietary/sensitive data into models and will bring the revised policy back to council for approval.
Ending: The council expects a revised policy at the next meeting, with new language clarifying permitted uses, prohibited datasets and oversight responsibilities.
