Library AI Policy
This policy guidance is intended to be used in the creation of an Artificial Intelligence (AI) policy for library staff. Below are foundational areas that we recommend are covered in an AI policy. Examples of policy excerpts are included in each section as inspiration for when building your policy. Each organization’s priorities are different, so select the components that resonate with your organization.
Core Components
Data Security and Confidentiality
AI may use data in ways that are not obvious or apparent to users. Names, contact information, borrowing history, or any other personally identifiable or proprietary information should not be entered into any AI tool without permission.
Policy examples: Customer privacy is a foundational library value. Staff shall not enter customer names, contact information, borrowing history, or any other personally identifiable information into any AI tool. All data must be anonymized before submitting.
Confidential information, trade secrets, or proprietary information may only be entered into AI tools if explicitly authorized in this policy or approved in writing by a supervisor.
Human in the Loop
Anytime information from AI is used, it should be reviewed and verified by the user. In some instances, AI responses may be information created by AI to satisfy a query. When using AI, staff should always review AI-generated content before using or distributing it. For more recommendations for staff training, please visit Getting Started with AI: Guide for Library Staff.
Policy examples: Staff shall review AI-generated content before use or distribution to ensure it complies with library policies and copyright law, does not disclose confidential information, and is accurate, current and unbiased. Staff are responsible for the content they use or act upon.
Staff shall not misrepresent AI-generated content as solely human-created. Disclosure of AI use should be exercised with professional judgment based on the context and audience.
Staff shall not rely on AI tools for legal advice or opinions.
AI should not be used alone to make employment decisions, including hiring, promotion, discipline, or performance evaluations.
Legal Compliance
Users of the AI tools should not use it in a way that may violate law. This is mostly aimed at malicious use of AI for illegal purposes, but users should also be aware of ways that AI use may inadvertently be illegal. An example of this is the use of AI transcription services that record conversations and transcribes them for the user. Eleven states in the U.S. are All-Party Consent states and generally require the consent of all parties before recording a private conversation, such as a Zoom call, private staff meeting, or a phone call. Using AI to record and summarize those conversations without consent may be considered illegal in those states.
Policy example: Staff shall not use AI to engage in or solicit illegal activities.
Additional Considerations
ADA Compliance and Benefits
AI can provide improved accessibility in many areas, such as real time captioning, speech to text transcription, image description and language translations, however, AI tools themselves are not always ADA compliant. Many AI tools are web based and may have issues with keyboard navigation, screen readers, color contrast, understanding speech from individuals with speech impairments, among other barriers. Library staff and administrators should consider the limitations of AI as it relates to ADA when implementing tools and ensure that other accommodations are offered.
Policy examples: The use of artificial intelligence does not diminish the library’s obligations under applicable accessibility laws, including the Americans with Disabilities Act and other applicable federal and state requirements.
AI tools used by the library are evaluated for accessibility and compatibility with assistive technologies. Staff should consider accessibility when selecting, implementing, or using AI systems.
The library shall maintain reasonable alternative methods for accessing services and information when AI-based systems are unavailable or inaccessible to an individual.
Bias Concerns
AI systems learn from large datasets that may contain historical biases, stereotypes, or underrepresentation of certain groups. As a result, AI can produce outputs that are unfair, inaccurate, or discriminatory even when no one intended that outcome. A library policy does not need to solve bias, but it should acknowledge the risk and establish safeguards.
Policy example: AI-generated content may contain inaccuracies, biases, stereotypes, or omissions. Staff using AI tools should exercise professional judgment and review outputs for fairness, accuracy, inclusiveness, and alignment with the library’s values before relying on or publishing AI-generated content.
Copyright
AI gathers information from many sources on the internet. Unbeknownst to the users, some of these sources may be copyright protected. Users should make a best effort not to use copyrighted information.
Policy example: Staff should respect copyright laws when generating content from AI and ensure the AI tools are using information from the public domain, not copyrighted sources.
Records Retention/Freedom of Information Act Requirements
Libraries should ensure that your internal practices are consistent with state and federal records retention and Freedom of Information Act requirements. Always refer to your legal counsel if you have any questions about what AI information would be required to be retained.
Policy example: AI prompts and outputs are subject to the same retention rules as other library records.
The use of AI does not create a new retention obligation. AI-related materials that are retained by the library may be subject to disclosure under the Illinois Freedom of Information Act (FOIA), consistent with applicable exemptions.
- Final records created using AI tools shall be retained in accordance with the library’s approved retention schedules.
- The library is not required to retain AI prompts unless they are incorporated into or relied upon as part of an official record.
Recommended Actions
Periodic Review of Policy
AI technology is changing at a rapid pace. It is recommended that AI policies, procedures or tools are reviewed periodically to ensure continued compliance with intended guidelines and usage.
Policy example: The Library shall periodically review this policy to ensure continued alignment with evolving artificial intelligence technologies, applicable laws and regulations, industry standards, and the library’s operational objectives.
Training
It is recommended that libraries provide or encourage periodic training to staff, so they understand the pros and cons of using artificial intelligence, and how to use it effectively. For more recommendations for staff training, please visit Getting Started with AI: Guide for Library Staff.
Policy examples: The library will provide periodic training to ensure staff understand appropriate AI use, including their responsibilities related to records retention, FOIA compliance and the handling of public records.
As AI technology evolves, so will our understanding and usage. We encourage staff to participate in workshops and training to become informed about new AI tools and techniques. Regularly assess the effectiveness of AI tools in your workflow. Sharing knowledge among staff can foster a collaborative environment and encourage innovative practices.
Violation Enforcement
Given the power of AI, it is very important that staff adhere to the guidance provided in the policy. If there is violation of the guidelines, disciplinary action should be outlined in the policy.
Policy example: Employees found to have violated this policy may be subject to disciplinary action, up to and including termination of employment.
Sample Policies
Arlington Heights (Illinois) Public Library (last updated April 2026)
Sayville (New York) Library (last updated May 2026)
Toronto (Canada) Public Library (last updated January 2025)
Last updated July 8, 2026
