AI & Libraries: Notes from the Field

The rapid advancement of artificial intelligence is prompting libraries to increasingly define their relationship with the technology. As I speak with library professionals in the US and abroad, there are some themes I see developing.

Administrators Want a Blueprint for Adoption!

Unfortunately, the road to emerging technologies in the library is often unpaved, and artificial intelligence is certainly no exception! More clarity is on the way in 2024, with a toolkit for U.S. educators, and pending AI legislation from the European Union. Despite the uncertain environment, AI avoidance is not a realistic plan, in part because:

Frontline Staff Want Direction from Administration Now!

Library workers often tell me they don't know if they're "allowed" to use AI, and if they are, in what circumstances! This lack of direction can lead to inconsistent service, and the potential for harm. While AI can be an incredible productivity tool, it also poses a serious threat to user privacy, if used inappropriately. I have had library staff describe personal use cases that range from innovative to highly problematic!

There is an Ethical Debate to be Had--and Now is the Time to Have it!

AI poses a number of ethical issues, ranging from its environmental impact to how training models are created, with respect to intellectual property. It is important to note that legal and ethical are two different matters; whatever legislation that may come could avoid sticky ethical questions, or resolve them in ways that may not align with our professional ethics.

In facilitating conversations on AI between library professionals, it is clear that feelings on these ethical matters are far from uniform. These debates can take time, but resolving them is necessary as AI implementation occurs. For example, is the use of AI-generated art acceptable within the institution, or is there an ethical standard that platforms are held to? The answers to questions such as these will impact what platforms the library invests in (if at all), and by extension, what tools it trains its staff on.

Library Workers Feel Unprepared to Evaluate Vendor AI Solutions

AI functionality is coming to library databases, chatbots, and other solutions. That said, many librarians that I have spoken to don't feel well-positioned to determine if a solution is good, bad, problematic, or even necessary. Safe experimentation with freely available AI tools (such as ChatGPT) will go a long way to understanding functionality, and establishing a baseline standard for potential adoption!

Concerns About Determining Whether Content is AI-Generated

In my travels, I hear phrases like "fake news on steroids," and "the death of truth". Uniformly, library workers recognize that AI, with its deepfakes, text-to-image generators, and other capabilities, poses a huge challenge to media literacy. To continue to confidently perform as information navigators, staff are seeking the tools and training necessary to determine if a piece of media they encounter is AI-generated. Non-profit organizations and educational institutions have begun to develop free resources for self-teaching, but library administrators need to make this subject a priority in staff continuing education.

AI detection tools are notoriously unreliable; the solution remains powered by people. In seminars I've conducted, we've focused on traditional media literacy skills, along with some new digital sleuthing techniques designed for the age of AI.

Until Next Time!

I hope you've found this article useful! Does your experience match my own? Below are some related items you may find of interest.

• March 21, 2024 — Virtual Panel Discussion: Using AI in the Library. Long Island Library Resource Council. Bellport, NY.
• Article: "Implementing AI in Libraries in Our Uncertain Environment". Republished at the-digital-librarian.com.