St. Louis Public Library recognizes that AI tools can expand access to knowledge while also acknowledging risks to security, privacy, transparency, and equity. This policy guides all St. Louis Public Library (SLPL) staff on the secure, responsible, ethical, and effective use of AI tools to enhance library services, improve efficiency, and spur innovation while protecting the privacy and rights of staff and customers. It applies to all SLPL staff and the use of AI tools in all library functions. This policy does not discuss customer use of AI, which is governed by the SLPL Technology and Appropriate Use policies.
SECURITY AND RELIABILITY
- SLPL Technology Services will create and maintain a list of SLPL approved AI tools for staff use. Staff must verify if a tool is SLPL approved before using it for work purposes.
- Because AI systems are prone to hallucinations (fabricating information), staff must critically evaluate and cross-reference AI outputs with reliable sources before use. AI should assist, not replace, professional judgment.
DATA PROTECTION AND PRIVACY
SLPL prioritizes data security and privacy. When using AI tools, all SLPL confidentiality and data privacy practices must be upheld.
- Staff is strictly prohibited from uploading or inputting Personally Identifiable Information (PII), customer information, or employee data into unapproved or Public AI systems.
- Uploading confidential or sensitive data is only permitted on SLPL approved AI systems when the data is necessary for the system's functionality, falls within the scope of its intended use, and the tool has been specifically approved for that purpose.
- Staff should log into approved AI tools using their organization-provided email address for all interactions related to SLPL to ensure data security and privacy.
RESPONSIBLE AND ETHICAL USE
- AI will not be used to deceive, manipulate, mislead, misrepresent, or otherwise falsify information.
- AI tools can enhance productivity, but do not replace professional judgment. Decisions impacting people’s lives should not be fully automated and should always be made by humans.
- Employees should not use AI to make employment decisions regarding hiring, termination, or promotions. AI should not be used to evaluate staff performance or set expectations.
- Staff should be mindful of the extreme resource needs and environmental impact of AI tools and consider using an alternative approach if the use of AI for the same task would yield a comparable result and use of staff time.
- Staff are accountable for the work produced using AI. They must check all AI-generated content for accuracy. If AI makes a mistake and it is not corrected, the staff member is responsible for the error.
-
When AI is used substantially in creating original works (e.g., blogs, social media posts, reports, marketing materials), acknowledgement of the role AI played is required. All such work must include a statement indicating the use of AI.
○ Attribution statements are not required when AI is used for tasks such as brainstorming or grammar and spell-checking.
BIAS AND EQUITY
The Library is committed to fair AI practices and prioritizes human-centered approaches.
● AI shall only be used in compliance with applicable federal, state, and local laws and regulations and Library policies.
● AI shall be evaluated for algorithmic bias and shall not be used to discriminate based on any protected characteristic under federal, state, or local laws or regulations.
COPYRIGHT AND INTELLECTUAL PROPERTY
- Use of AI and Machine Learning tools must not violate applicable laws, including copyright, trademark, or intellectual property rights, and must comply with all Library policies.
- Employees must closely review any output generated by AI for potential violations of copyright law.
TRAINING
The Library will provide ongoing training for staff on the use of AI in the workplace, including ethical use, bias recognition, privacy protection, and assisting customers with AI tools.
DEFINITIONS
- Artificial Intelligence (AI): Computer systems designed to perform tasks typically requiring human intelligence.
- Machine Learning: A branch of artificial intelligence focused on building systems that learn from data to identify patterns and make decisions with minimal human intervention.
- Generative AI: AI systems capable of generating new content (text, images, audio) based on user prompts.
- Public AI: Open platforms where input data may be used to train the model (e.g., ChatGPT, Midjourney). Staff should assume that data entered into public AI tools becomes part of a public dataset.
- Private AI: Systems operating in a closed environment where data inputs are restricted and not used for external model training.
- AI Hallucinations: The tendency of AI tools to fabricate information or present inaccurate facts confidently.