Like the fog in Carl Sandburg's poem, AI technologies arrive "on little cat feet". Unlike the fog in the poem, AI doesn't move on, dissipating as the day progresses and grows warmer, unless you count moving into other areas of your work, your institution, and even your life as "moving on".
Opinions about AI differ dramatically. AI makes life so much easier. AI is destroying the planet. AI is a wonderful co-creator and co-researcher. AI is extinguishing critical thinking. AI will lead to super intelligence. AI is dumbing down humankind.
Even within our information professional ranks, no absolute agreement on the benefits and harms of AI exists. We are, however, more likely to be aware that AI encompasses multiple technologies and is not one, simple, stand-alone entity. Artificial (sometimes called Augmented) intelligence has been in existence for decades. Only in the last few years, with the introduction of generative and agentic AI into general usage, have opinions heated up. No one worried about the demise of humanity when using spell check.
In response to the fear of AI by some library patrons, several public libraries in the U.S. have begun to offer AI avoidance programs. One of the first to offer avoidance workshops was at the public library in Bangor, Maine. Other librarians heard of the program and started their own. In South Philadelphia, Pennsylvania's public library, the emphasis is on removing unwanted AI from apps on patrons' phones.
These programs caught the attention of the non-library press. For coverage of them, read these articles in TechCrunch and Fast Company. Remio, a company selling AI agent technology, also weighed in on the viral nature of the library workshops, as did India-based Kalinga.
Amanda Siberling, in the TechCrunch article, makes the point that attendees at the avoiding AI workshops aren't what we might consider as Luddites. She writes, "the anti-AI movement isn't about rejecting technology altogether so much as it is about advocating for more control, agency, and freedom in how people use technology."
Not all libraries have explicitly adopted AI avoidance, preferring to concentrate their efforts on AI literacy. The thinking goes that, once people understand how AI works, where it appears, and how it affects privacy, they will be better equipped to decide the pros and cons and whether to either embrace or circumvent individual aspects of AI technologies.
Academic libraries have generally taken a different approach to AI than promoting avoidance. They worry about the effect of generative AI on learning and on research. Libraries worldwide have instituted programs explaining the ethical and responsible use of AI to students. Critical AI literacy programs abound, with some academic institutions requiring students to adhere to AI usage policies.
It's not only individual institutions with AI policies. IFLA published a "Statement on Libraries and Artificial Intelligence" in 2020 and "Guidelines on AI in Libraries, 2023-2024" in 2025.
As vendors explore the benefits of AI, it becomes harder for librarians to distinguish between what is happening with our subscribed information sources and what is occurring in the largely ungoverned world of frontier AI models created in the labs of companies such as OpenAI, Meta, and Anthropic. Two developments of interest are the Elicit Research Agent, introduced in December 2025, which uses agentic AI to alert researchers to new information relevant to their work. The National Library of Medicine, without much fanfare, introduced Linked Discoveries in PubMed as a pilot project. It uses to linked data to surface how medical literature is connected. Libraries are also seeing new mechanisms for database searching from Clarivate, Factiva, LexisNexis, and many smaller companies.
Although opinions about AI run the gamut from complete feat to unbridled enthusiasm, there is little doubt that the technology is here to stay. How it is implemented and what elements of it librarians will adopt is still under consideration. Total avoidance of AI is not really an option in today's world. Let's hope that human agency will remain in the many iterations of AI that have appeared and will continue to appear as time goes on.