Kootenai County staff review draft AI use policy; IT to distribute to elected officials
Loading...
Summary
Kootenai County IT presented a draft policy guiding employee use of generative artificial intelligence, emphasizing supervisor approval before entering county data and cautioning about AI inaccuracies; commissioners directed distribution to elected officials and discussed next steps toward formal adoption.
Kootenai County IT Director James Martin presented a draft policy on June 18, 2025, that would guide county employees’ use of generative artificial intelligence tools, saying the policy requires employees to check with supervisors before entering county data into AI services and calls for transparency about AI use in work products.
The policy, Martin said, was developed by ICRMP and tailored for Kootenai County. "This policy applies to all staff who desire to utilize AI technology to assist them in fulfilling their job functions. All employees of Kootenai County are expected to interact with generative AI technologies in a responsible and ethical manner," Martin read aloud during the Human Resources meeting.
Commissioners and staff put forward two practical aims: reduce the risk of sensitive county data being submitted to third-party generative systems, and give supervisors visibility into how employees are using AI. Martin said IT would monitor developments in the technology and update the policy as needed. He noted the county’s principal concern is the entry of sensitive data into third-party AI tools.
During discussion, commissioners asked whether employees must obtain written permission before using AI and whether IT must be notified when staff use web-based tools versus downloaded apps. Martin said the draft requires employees to check with their supervisors before entering county data and that IT would like to be notified so it can be aware of tools in use; he also noted that many generative tools are immediately accessible on the web but that downloaded applications raise additional network considerations.
A commissioner urged caution about relying on AI-generated text because of so-called "hallucinations"—instances where AI produces inaccurate or fabricated material. Martin said the draft reflects that concern and requires employees to verify AI-produced content and disclose when AI assisted in producing work products.
Next steps discussed included distributing the draft to elected officials for feedback and placing the policy on a future business meeting agenda for adoption. Martin said he would distribute the draft to elected officials following the HR meeting and gather any feedback before moving the item to a business meeting.
The board did not vote on the policy during the HR meeting and directed staff to collect elected-official feedback and return with any recommended revisions prior to formal adoption.

