ASHLAND — Artificial intelligence is moving quickly into workplaces across the country.
At the Ashland County Health Department, officials are trying to make sure its use doesn’t move faster than the rules governing it.
The Ashland County Board of Health unanimously approved a new AI governance policy Tuesday. It establishes rules for how employees can use the technology while conducting business.
The policy applies to employees, contractors, interns, volunteers, board members and other affiliated personnel using AI for agency business.
Health Commissioner Vickie Taylor said the department also placed a moratorium on most AI use until employees complete training on the new policy Monday. Employees can still use AI for limited purposes with approval from their supervisors, she said.
The policy allows AI to assist with tasks such as drafting, summarizing, organizing and analyzing information. However, it states that AI cannot independently make agency decisions or replace professional judgment.
Staff also remain responsible for reviewing AI-generated work.
“We do want to make sure that human element is in there,” Taylor said. “That you’re looking to make sure that it’s hitting health equity, that it’s accurate and the sources are accurate.”
Below is a copy of the Ashland County Health Department’s AI governance policy. The story continues below.
Policy: AI-generated content will need reviewed
Under the policy, AI-generated content must be reviewed, fact-checked, corrected and approved by qualified staff before it is treated as authoritative.
The policy also prohibits employees from entering protected health information, personally identifiable information and other sensitive agency data into unapproved AI tools. That includes confidential public health data, personnel records, investigative records, legal materials, credentials and security information.
The department will also require oversight of AI-assisted work when the technology materially influences or changes a work product, decision, communication, analysis or action.
Minor uses such as spelling corrections, grammar suggestions, formatting assistance, transcription and basic scheduling support generally do not fall under that definition.
Taylor said she spent significant time developing the policy because she wants the department to understand how employees are using AI and ensure they follow established protocols.
“I just don’t want to get caught into a situation where we just let it go rampant and then something happens and then we have no documentation or policy,” Taylor said.
The policy also directs the department to consider whether AI tools could produce biased or inequitable results and whether they are accessible and appropriate for vulnerable or underserved populations.
It additionally recognizes the environmental impact of artificial intelligence, including the energy and water used by data centers that train and operate AI systems.
The department’s leadership team will currently oversee the agency’s AI governance, including the approval of AI tools and review of emerging risks, vendor changes and policy updates.
