10.  District Operations and Administrative Policies

10.1.               Artificial Intelligence (AI) Use Policy

10.1.1. Policy Purpose

10.1.1.1.      The purpose of this policy is to establish clear guidelines meant to encourage purposeful and responsible use of generative artificial intelligence (AI). This policy aims to foster public trust, support business outcomes, and ensure the ethical, transparent, accountable, and responsible implementation of this technology. The use of AI, including generative AI, has the ability to expand creativity, produce quality content, and enable the quick gathering and analysis of data to improve the delivery of District services. AI must be applied with great care because it has the potential of perpetuating security risks, among many other potential negative impacts. A careful, prudent, and well-planned AI policy may protect against misuse, drive innovation, support increased efficiencies in operations, and improve experiences for community engagement.

10.1.2. Departments Affected

10.1.2.1.      All departments are affected by this policy.

10.1.3. References

10.1.3.1.      This policy draws reference from the following: a. Interim Guidelines for Purposeful and Responsible Use of Generative Artificial Intelligence (AI) in Washington State Government. b. Public Records Act – RCW.42.56. c. Criminal Justice Information Services (CJIS) Security Policy.

10.1.4. Definitions

10.1.4.1.      AI (Artificial Intelligence) – Refers to the simulation of human intelligence in machines that are designed to think and act like humans, including learning, reasoning, problem-solving, perception, and language understanding.

10.1.4.2.      Generative AI – A subset of AI that focuses on generating new content, such as text, images, or music, based on existing data. It uses algorithms to create new data patterns and outputs that mimic human creativity and can create content (including text, images, audio, or video) when prompted by a user.

10.1.4.3.      GPT – Models trained on large datasets to understand and generate human-like text. These models are pre-trained on a diverse range of internet text and can perform various language tasks with minimal fine-tuning (e.g., ChatGPT, Anthropic, Gemini, Copilot).

10.1.4.4.      Confidential / Non-public Data – Confidential information that is protected from disclosure or release by law. Examples include: social security numbers, a driver’s license number or a Washington identification card number, account numbers (e.g., utility account), credit card numbers, security codes, or passwords. Additionally, this includes data held in personnel records, such as residential phone numbers and addresses, personal cell phone numbers, personal addresses, and emergency contact information.

10.1.4.5.      Publicly Accessible Generative AI – An AI system that is not isolated to or owned by the District, but rather is available for any person to access and use via the internet.

10.1.4.6.      Workflow – A sequence of tasks or activities designed to achieve a specific goal or outcome, often involving multiple steps and individuals.

10.1.5. Usage Guidelines

10.1.5.1.      Records Safety - Integrating, entering or otherwise incorporating Confidential / Non-public Data into publicly accessible generative AI or GPT systems is prohibited. The use of such data could lead to unauthorized disclosures and legal liability.  State law restricts the sharing of Confidential / Non-public Data with unauthorized third parties, including publicly accessible generative AI or GPT systems.  Using a generative AI system may result in creating a public record under Washington state's Public Records Act.

10.1.5.2.      Acceptable Use - Leveraging AI to push a user's workflow along, spark innovation, or seed ideas is encouraged, as long as Confidential / Non-public Data are not used or inputted.

10.1.5.3.      Guidelines - The Board of Commissioners and Staff must adhere to the following guidelines when using AI:

10.1.5.3.1.          Establish clear goals for the particular use case.

10.1.5.3.2.          Consider the potential impact of generative AI on stakeholders, including customers, employees, vendors, contractors, and partners. 

10.1.5.3.3.          Regularly review and audit the generative AI output to ensure it meets quality standards.

10.1.5.4.      Considerations - When using AI, the Board of Commissioners and Staff must be mindful of the following limitations and requirements:

10.1.5.4.1.          Generative AI systems can fabricate answers, citations, and articles. None of the output should be trusted without verification and validation of the answers.

10.1.5.4.2.          The knowledge base or AI output data may not be the latest information available, as many AI language models may be more than a year old.

10.1.5.4.3.          Usage of generative AI to draft any record, formal document, or publishable document item must include a thorough manual check of all data as part of the workflow.

10.1.5.4.4.          The Board of Commissioners and Staff should not assume the copyright or intellectual property status of any generative AI system output. Terms of service, end-user license agreements, and other similar documentation may add restrictions to the usage of the output.

10.1.5.4.5.          If entering District-specific data, the Board of Commissioners and Staff should ensure the terms of use include policies around data retention and protection. If uncertain, they should contact the District’s Legal Counsel for guidance. Any District data retained by the platform should be purged upon termination. Additionally, no District data should be sold or transferred to 3rd parties.

10.1.5.4.6.          All Technology purchases should be coordinated through the General Manager.  Paid subscriptions to generative AI platforms may offer more favorable terms and conditions and are generally preferred over free/public versions.

Adoption Date: November 20, 2025