Loading...
HomeMy WebLinkAbout100-25POLICY & PROCEDURE Subject: ARTIFICIAL INTELLIGENCE (AI) POLICY Index: Executive Services Department – IT Number: 100-25 Effective Date: June 16, 2026 Supersedes: None Page: 1 of 6 Staff Contact: Deputy Chief Administrative Officer Approved By: 5/11/2026 1.0 PURPOSE: The City of Renton is committed to the ethical, transparent, and legally compliant use of artificial intelligence (AI) to improve services, operations, and decision-making. This policy establishes high-level expectations for the responsible use of AI, with detailed procedures maintained by individual Departments to separately guide implementation and best practices. 2.0 SCOPE: This Policy is applicable to all city of Renton employees, public officials, and volunteers of the City who use or desire to use AI assisted tools to generate or modify work product for the City (“AI Users”). 3.0 RELATED REFERENCES: Public Records Act – 100-05 Electronic and Information Technology, Management and Usage – 100-21 Electronic Data Security – 100-22 Procurement Requirements – 250-02 Discipline – 300-47 4.0 POLICY: 4.1 Review and Approval of AI Tools: 4.1.1 Except as provided herein, any usage of AI tools to produce work product (whether those tools are paid, free, or installed on City devices as software or accessed through internet/cloud-based services) must be reviewed and approved for the intended use by the user’s Department Administrator, or designee. Artificial Intelligence (AI) Policy Policy 100-25 p. 2 Department level approval is not required for use of AI tools that are incorporated within City IT approved software and perform only basic, low risk assistive features such as spell check, grammar check, sorting, and summarization or simple formatting that do not create original work product or substantially alter human generated original work product. 4.1.2 In addition to Department level approval, any uses that involve a) uploading/sharing Potentially Sensitive Information with the AI tools, b) installation on City devices, or c) paid services must be reviewed and approved pursuant to city policy 250-02. 4.2 Accountability and Disclosure: AI users are responsible for the accuracy, appropriateness, and impact of any AI-generated content they use to make decisions or incorporate into work product. Any user that incorporates and shares AI generated content that they have not personally verified as accurate must label or identify the extent to which unverified AI content is used, so others can either verify for themselves or avoid relying on the content. 4.3 Confidential Data Protection: Most AI tools are not secure or appropriate for confidential use. Potentially Sensitive Information may only be entered or uploaded into AI tools that are designed to safeguard the particular type of confidential data – meeting any required legal, contractual, compliance, or policy requirements to protect its confidentiality. Before use, such safeguards shall be documented by contract and confirmed by the IT division as safe to use for the intended purpose. Users of the service are responsible for understanding any limitations in what information can be shared with the AI tool, and otherwise conform to all requirements of this Policy. 4.5 Prohibited uses: 4.5.1. AI users shall not use AI generated content to make decisions they do not have the time, qualifications, or expertise to reach without the use of AI. For example, non-attorneys should not rely upon legal answers by AI, users without coding experience should not rely upon code generated by AI, and non-engineers should not rely upon engineering analysis performed by AI. 4.5.2. No user of AI may use AI in a way that creates an unlawful discriminatory impact on a person based upon their legally protected characteristics. Artificial Intelligence (AI) Policy Policy 100-25 p. 3 4.5.3. No user of AI may input Potentially Sensitive Information into an AI tool that has not been approved for such use or that would violate an applicable law, contract, or policy. 4.5.4. No user of AI may use or share AI generated content without subject matter expert review and oversight, or disclosure that the content has not been verified for accuracy and should not be relied upon. 4.5.5. No user of AI may use or distribute AI generated content or tools such as pictures, video, human voices, or chat bots that could be interpreted by a recipient/user as being or depicting real people unless the use of AI is clearly labeled and disclosed. 5.0 DEFINITIONS OF KEY TERMS: 5. 1 AI (Artificial Intelligence): The capability of a machine or system to perform tasks that typically require human intelligence, such as generating text or audiovisual content, making or recommending decisions, analyzing data, or automating processes. 5.2 AI Tools: Software applications, platforms, or systems that use AI technologies to perform tasks. These tools may include, but are not limited to, solutions for text generation, image or video creation, predictive analytics, data classification, conversational interfaces, and process automation. 5.3 Potentially Sensitive Information: Includes all information that is protected by law, contract, or policy. This includes, but is not limited to: any information that would be redactable under the Washington State Public Records Act (Chapter 42.56 RCW), attorney-client communications and work product prepared in anticipation of litigation, Criminal Justice Information (CJI), utility billing and customer account information, Taxpayer or financial information submitted to the City, medical records and health-related information that is protected by law, information for which release could violate privacy rights or other legal protection. 6.0 PROCEDURES: 6.1 AI Tool Use and Approval 6.1.1 Prior to acquiring AI Tools or AI supported services, employees responsible for selection of AI tools and services shall work with an IT Artificial Intelligence (AI) Policy Policy 100-25 p. 4 representative to evaluate any AI components of the services to ensure they can be implemented for the intended use consistent with this Policy. This applies to both software-based services and professional or other services. 6.1.2 All AI tools must be approved prior to use for City purposes by the users’ Department Administrator, or designee. This approval requirement applies to the use of both free and paid tools (e.g. ChatGPT, Perplexity, Gemini, Copilot, ect) regardless of whether the service is installed on a City owned device or accessed using city or personally owned devices. 6.1.3 Use of AI tools involving Potentially Sensitive Information, paid services, or installation on City devices must follow the review process outlined in the policy or separately published guidelines. IT, and other subject matter experts as appropriate, will review the AI service for compliance with applicable legal, policy, security, and contractual requirements. IT will also maintain a list of approved AI tools that identifies types of Potentially Sensitive Information that can or can’t be used with such services. 6.1.4 Regardless of IT approval and existence on a list of IT approved AI tools, users remain responsible for complying with all aspects of this Policy and obtaining supervisor approval for how such services may be used for their specific roles. 6.1.5 Additional training or guidance documents may be created and enforced at the city-wide, department, division, or team level to ensure guidance and restrictions focus on the type of work being performed with AI assistance. 6.2 Use of Potentially Sensitive Information: 6.2.1 Users of AI tools may not input Potentially Sensitive Information into any AI tool unless the system has been approved for use with that specific type of data by IT and their Department Administrator. 6.2.2 Approval must include verification of contractual and technical safeguards appropriate to the type of information involved, and identification of potential bias or tool limitations that would make use of the tool inappropriate or require specific training before use. 6.2.3 Users are responsible for understanding the limitations of the AI tool, ensuring compliance with applicable requirements of this policy. Artificial Intelligence (AI) Policy Policy 100-25 p. 5 Depending upon the use case, training should be obtained as necessary to properly use AI tools in an ethical and responsible way. 6.3 Records Management and Public Records Compliance: 6.3.1 AI-generated content must be managed in accordance with the Washington State Local Government Common Records Retention Schedules (CORE) and applicable City policies. 6.3.2 If AI-generated content is to be retained for more than transitory purposes, it must be stored in a City-managed system and be retrievable until properly disposed of for City use and responding to public records requests. Transitory records that are created on systems for which other staff members in the City do not have access should be deleted when no longer needed as long as allowed by CORE and not subject to a pending public records request, litigation hold, or other legal requirement to retain. 6.3.3 Systems that are not centrally managed/accessible by others in the City should not be used to create records with more than transitory retention value, and settings for retention of records in such systems shall be set to as short as period as allowed by law and policy and/or manually deleted as soon as no longer needed. If records with retention value are created in such systems, users shall download and preserve the records on City managed systems and maintain in accordance with retention schedules. Any records available only to an end user that remain at the end of their employment or term of service for the City shall be either deleted in accordance with records retention schedules or downloaded and preserved on City managed resources. 6.3.4 If records are created within AI tools in a non-centrally managed system that is accessible and searched by designated public records staff, users of AI tools are responsible for searching for and producing all public records as necessary to respond to public records requests and/or requests to comply with legal requirements. Such users are also responsible for preserving records that are subject to request in order to comply with legal holds and public records request related preservation requirements. 6.4 Training 6.4.1 When new uses of AI tools are approved, the approvers (IT and Department Administrator, or Designee) should identify any required Artificial Intelligence (AI) Policy Policy 100-25 p. 6 training for the responsible and ethical use of the tool. The type and extent of training may vary based upon whether the use is a high-risk use (e.g. involves consequential decisions, Potentially Sensitive Information, and/or impacts legal or other rights of other people).