Chelan’s HR Director, Chad Coltman, presents a draft policy covering City use of AI in the workplace to the City Council on Oct. 27.
ANDREW SIMPSON/WARD MEDIA
Posted
Saturday,October 31, 20262:48 pm
CHELAN — City employees could use artificial intelligence to draft reports, research projects and streamline daily work under a proposed policy that would also require them to verify AI-generated information, protect confidential records and disclose when publicly distributed images or videos were created using the technology.
The Chelan City Council received its first look at the proposed artificial intelligence use policy Oct. 27, as city officials seek to establish guidelines for technology that is increasingly incorporated into everyday workplace software.
HR and Communications Director Chad Coltman said the goal is not to discourage employees from using AI, but to establish consistent expectations for how they use it and remain responsible for the results.
The proposal would apply to city employees, elected officials, volunteers, contractors and others authorized to access city information, including when they conduct city business on personal devices.
Chelan’s preferred platform would be Microsoft Copilot, which integrates with software already used by city employees, including Outlook, Word and Excel.
Under the proposed policy, employees could use the free version of Copilot for research, brainstorming and drafting based on publicly available information. Licensed enterprise versions, which provide additional security and administrative controls, could also be used with authorized city business information.
For more complex tasks, including technical analysis and extensive document preparation, the city would favor an enterprise version of Claude, another AI platform.
Personal accounts on services such as ChatGPT and Google’s Gemini would generally not be permitted for official city work without prior authorization.
Coltman explained that the distinction is largely about protecting information. Publicly available material can be used in approved free tools, while internal documents and sensitive information require additional safeguards.
The draft policy would prohibit employees from entering confidential city information into unauthorized AI systems. Sensitive personal information, including Social Security numbers, financial account details and medical records, would be restricted to approved enterprise systems under appropriate safeguards.
Attorney-client communications, executive session materials and certain law-enforcement-sensitive information would require specific authorization from the city attorney before being entered into an AI system.
The policy also establishes limits on how AI can influence government decisions.
AI could assist staff in researching regulations or preparing reports, but could not serve as the sole basis for permitting, enforcement or other government decisions affecting residents. The same restriction would apply to employee hiring, discipline, promotion and termination decisions.
Human review would be required before any AI-generated work is released or relied upon.
Coltman noted that AI systems can produce inaccurate information, sometimes referred to as hallucinations, and that employees must verify facts, figures and sources before using the results in official work.
The proposed policy also addresses public records.
Prompts, conversations and documents generated through AI tools may qualify as public records when they relate to city business. Employees would be required to retain qualifying records in approved city systems and could not assume that an AI conversation is exempt from disclosure simply because it occurred through a chatbot.
Another provision would require clear labeling of AI-generated or AI-assisted images, audio and video used in public-facing materials, including city websites, social media, presentations and printed documents.
Coltman said the disclosure requirement is intended to prevent confusion about whether an image or recording represents something real.
During council discussion, members asked how the restrictions would affect ordinary research, including gathering information from internet searches that increasingly incorporate AI-generated responses.
Coltman said the city does not want to make its policy so restrictive that employees cannot use common research tools. For example, a simple Google search now returns results gathered through Gemini, so if the policy were too restrictive, employees literally couldn’t “just Google it” when looking for an answer to something. Instead, staff would be expected to verify information gathered through those tools and identify sources when appropriate.
The proposal would also prohibit AI use for political campaigns, lobbying or partisan activities conducted through city systems.
If adopted, the policy would require employee training on approved tools, information security, public records responsibilities and appropriate use. Violations could result in retraining, loss of access to AI tools or disciplinary action, depending on their severity.
The city administrator would oversee the policy and approve AI tools, with the city’s information technology staff responsible for security reviews and account management.
The draft calls for the policy to be reviewed at least annually as the technology and applicable regulations evolve.
Council took no action Oct. 27. The policy is expected to return for additional review before members consider whether to adopt it.
Andrew Simpson: 509-433-7626 or andrew@ward.media
