Pinal County employees now have clearer rules for using artificial intelligence at work. On March 18, the Board of Supervisors approved Policy 2.61, a new framework governing the responsible use of AI tools in county operations. The policy lists an effective date of March 18, 2026.
What the Pinal County AI Policy Covers
Policy 2.61 establishes a formal list of approved AI tools, maintained by the county IT department on the employee intranet. Employees may use approved tools for authorized purposes including:
- Analyzing data
- Drafting emails, letters, memos, and presentations
- Creating outlines and checklists
- Conducting research
- Assisting with repetitive tasks
- Data processing
However, the policy draws firm limits. Employees are prohibited from using AI to create synthetic images, videos, or audio of real people without consent; create election materials without the consent of the Communications department, Officer in Charge of Elections, and the County Recorder; make HR decisions; create legal documentation; or share confidential or restricted data. The policy also prohibits knowingly creating false information, attempting to identify individuals from photo or audio, and bypassing security restrictions.
The policy also states that free AI systems are not approved.
Google Gemini: The Primary AI Tool for County Staff
Jason Cantrell, Pinal County IT CIO, highlighted Google Gemini as the AI tool expected to benefit the broadest range of county employees. Holley said it can be integrated into the county’s Gmail system and that employees can optionally enable it to assist with tasks such as email drafting and grammar.
Chairman Jeff McClure pointed to responding to high email volume and producing presentations as immediate use cases. Cantrell agreed. “Gemini is the overall that I feel will be the biggest advantage to the everyday employee in improving efficiencies,” he said. He described applications such as taking data points, combining them, and generating outputs — as well as building presentations.
Cantrell said the county has approved a few other tools for specific departmental use cases, and that additional options are under review.
AI in Plan Review: A Future Possibility
Supervisor Stephen Miller raised the prospect of using AI to accelerate plan review. He described a scenario where AI could compress an 85-page submission down to a list of deficiencies, allowing staff to spot-check results or review every item individually. “I see it as a tool not to replace people, but to compress the time,” he said.
Cantrell said the options have not been approved or purchased yet, but confirmed the county is exploring them. “We will be going down that road to see if it’s available for Pinal County,” he said.
April Holley, who presented the policy, described the intended role of AI in that process. “It’s not meant to replace a person. It’s to give us an output that we can then review as humans to make sure it’s actually taken everything we wanted it to into consideration, making sure everything is correct — helping us get to the end result faster. The important part is that afterwards, there is human review.”
Data Privacy and Security Requirements
The policy requires employees to comply with applicable privacy laws when using AI tools, including HIPAA and Arizona Criminal Justice Information System (ACJIS) standards. Staff may not enter any Pinal County, employee, customer, or third-party confidential, personal, or proprietary information into an AI prompt without prior written authorization from the IT department.
Holley noted during the meeting that health information should not be inappropriately shared or accessed through unauthorized AI tools. The policy addresses that risk directly by restricting what data can be entered into any AI system.
The policy also notes that while historical inquiries and prompts using Google Gemini are not retained by county IT, they may still be subject to public records requests under Arizona law. For that reason, the policy calls on employees to write prompts with the same professionalism they would apply to any public-facing communication.
Ethical Use and Human Oversight Requirements
The policy treats all AI-generated content as a draft requiring human review before use or distribution. Under its terms, staff are expected to check outputs for biased, offensive, or discriminatory content, and to verify factual accuracy before sharing them inside or outside the county. AI tools may not be cited as authoritative sources and cannot substitute for legal, policy, or technical expertise. The policy also bars critical decisions — involving employment, customer relationships, finances, or business strategy — from relying solely on AI recommendations.
Where a commercial AI model is used, the policy requires employees to document which tool, prompts, and methods they employed. The policy includes suggested disclosure language noting that content was AI-assisted and subsequently revised by county staff.
Elections Department: Explicit Restrictions
The policy sets firm boundaries for the Elections Department specifically. According to the policy, AI tools may not interact with certified elections equipment — including the Election Management System, tabulators, and expressvotes — which are certified by the Election Assistance Commission and the Arizona Secretary of State and cannot be modified.
Enforcement and Compliance
The policy gives the IT department authority to audit any county department’s AI tool usage for compliance. Employees found in violation face disciplinary action up to and including termination.
Employee Communications and Training Schedule
Holley said the Gemini integration with the county’s Gmail system will be optional for employees.
The item passed with no opposition recorded.
