I recently had the chance to tune into the AI Policy in Libraries: Firsthand, Frontline Examples webinar hosted by Bibliotheca. Featured speakers Shelby Moffatt (Community Engagement Specialist at Whitby Public Library) and Adam Haigh (Technology Librarian at Lower Merion Library System) shared some really grounded, practical insights into how public libraries are navigating artificial intelligence right now — not in theory, but on the ground. Here’s what I took away from their conversation.
Why a library needs an explicit AI policy
The panel made a strong case that an AI policy isn’t just about staying current — it’s about protecting your community and your staff. A few key reasons stood out:
- Setting boundaries for patron expectations. Libraries are primary community tech hubs. A clear policy establishes how staff can assist patrons with AI tools while setting realistic expectations about what the library can and cannot support.
- Protecting privacy and intellectual property. Without clear guidelines, staff or patrons could accidentally enter sensitive information — patron data, internal records, copyrighted works — into public AI platforms. That’s a risk no library wants to take.
- Addressing ethics and misinformation. Policies help institutions proactively address algorithmic bias, AI “hallucinations” (those erroneous outputs AI is known for), and proper attribution for AI-generated content.
Two paths to policy formulation
One size does not fit all, and the webinar contrasted two primary approaches nicely:
- Standalone policies — writing a dedicated policy created specifically for AI (the route Whitby Public Library took).
- Integrated guidelines — updating existing acceptable use or computer technology policies to incorporate AI-specific language (the approach Lower Merion Library System used).
Both have merit, and the choice really depends on your library’s structure and culture.
A few technical terms, simplified
The discussion touched on some industry concepts worth breaking down:
- Agentic AI — rather than just responding to single prompts, “agentic” systems can independently perform complex, multi-step tasks (think generating a report, creating slides, and drafting follow-up emails in sequence).
- AI superuser — a team member who actively uses AI tools to streamline repetitive tasks, speed up administrative workflows, and boost creative projects. Not a formal title, but a useful role to identify on your staff.
- Standalone vs. integrated policy — think of a standalone policy as a brand-new handbook, whereas an integrated policy adds a new chapter to an existing rulebook.
Practical steps for library administrators
If you’re looking to implement or update AI guidelines at your organization, here’s the roadmap the speakers laid out:
- Audit existing guidelines. Determine whether an integrated update or a standalone document makes the most sense for your library’s structure.
- Involve frontline staff early. Gather feedback from technology librarians, reference staff, and community specialists who interact directly with patrons and daily operations.
- Define data guardrails. Clearly outline what data — especially personally identifiable information — should never be typed into public generative AI tools.
- Foster staff confidence through experimentation. Provide safe spaces, training, and simple use cases so staff can build familiarity without fear.
- Treat policies as living documents. Because AI technology advances so quickly, schedule periodic reviews — every 6 to 12 months — to update your language and rules.
A special thanks to Bibliotheca for hosting this timely conversation and providing practical resources for the library community. If you missed the live session, I’d recommend keeping an eye out for the recording — it’s well worth your time.
