Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Technology Policy topic
No spam. Unsubscribe anytime.
Kenosha County adopts generative AI policy establishing training, logging and authorization safeguards
Summary
County IT presented and the committee approved a generative AI usage policy requiring human review, data-protection measures, authorized-user training and logging of prompts; staff said pilot approaches include Microsoft Copilot and exploring local models to reduce data exposure risk.
Kenosha County IT introduced a generative artificial intelligence policy to the Finance & Administration Committee on April 16, and the committee approved the policy.
Mike, the county’s IT lead, said the policy is intended to allow responsible use of AI while protecting sensitive data and ensuring employee accountability. The policy requires that employees use unbiased documentation as inputs, that they verify AI outputs, and that employees be trained and authorized before using AI on county equipment. Mike emphasized that uploading sensitive documents to free third-party tools can expose data and that the county must control tools and logging where possible.
Committee members asked how staff would check AI hallucinations and whether local models were practical. Mike said verification is primarily the responsibility of the employee using the tool and that logging and prompt capture enable after-the-fact auditing; he also described vendor demonstrations of locally hosted models (for instance, a Mac Mini loaded with a Gemini-class model) that could run county data without sending it to third-party services.
Supervisors asked how ‘‘authorized personnel’’ would be identified and trained. Mike said the county plans a phased approach: basic guidance and training materials for broad awareness, then targeted training sessions and license assignment only after training. He said that some county staff already have limited access to enterprise tools such as Microsoft Copilot and that formal licenses would be given to users after training.
The committee approved the AI usage policy by voice vote. Mike said follow-up will include developing training modules, establishing logging and audit procedures and defining the authorization workflow.
Ending: The policy sets a framework for responsible AI use in county operations; staff will return with training materials and procedural details for audit and authorization.
