Yes, ZERO.
That is the number of libraries providing paid AI tools to their own staff.
Not low adoption. Not a slow rollout. Zero.
I ran a quick poll in the AI4LIB community last week with one question: Do you have access to a paid LLM (ChatGPT, Claude, Gemini, Perplexity), and if yes, who pays for it? Out of 24 respondents:
14 have zero access to paid AI tools. 7 pay entirely out of their own pockets. 3 have personal & institutional access. 0 work at a library that budgets for staff AI seats.
I read that number three times before it sank in.
For two years, our field has published papers, run webinars, and written policy documents on how AI is transforming library services. We draft policies on ethics, disclosures, and patron opt-outs.
Meanwhile, actual librarians are quietly spending around ₹400/month from their own salaries just to keep up—because their institution hasn’t budgeted a single seat for them.
You cannot teach AI literacy to patrons using tools you were never given.
The Budget Disconnect When library administrations do allocate funds for AI, they usually ask the wrong question:
“Which AI tool should we buy for our patrons?” (Result: Paying thousands for vendor AI tools nobody opens.)
The right question is:
“What should we give away for free before spending a single rupee?”
The Practical AI Hierarchy for Libraries Before renewing a single AI subscription this year, prioritize your tech stack like this:
1. Fund Your Staff First Give your team access to at least one paid reasoning model. If there’s no line item in your budget for staff tools, you aren’t behind the curve—you are the curve.
2. Maximize the Free Core (Start Here)
Semantic Scholar: AI summaries across 200M+ papers. Should be bookmarked at every reference desk.
NotebookLM: Upload PDFs for instant Q&A, summaries, and audio explainers. Replaces half your orientation deck.
ResearchRabbit: Visual citation mapping for discovery. Still the best-kept secret in academic libraries.
3. Audit Existing Contracts
Half of “buying AI” is reading your current subscriptions. EBSCO natural language search, Libby features—check what you already own. 4. Buy Paid Tools Only on Demand
Elicit: For grad students running systematic reviews.
Scite: For scholars checking if claims hold up under citation scrutiny.
R Discovery / Paperguide: Useful if you want one platform instead of five separate logins.
The 6-Month Test Before your next procurement meeting, ask one honest question: Did a patron ask for this paid AI tool by name in the last 6 months? If the answer is no, you already know what to do with that budget.
I want to hear it plainly in the comments (I’m counting):
Does your library pay for any AI tool for its staff?
What’s sitting on your library’s subscription list that nobody actually uses?