Personas of librarians and their response to AI - part 2

To see part 1 -

Group 6: The IL Hype Merchant

"We need to teach AI literacy—I've developed a framework."

First to jump on the bandwagon teaching prompt engineering. Rushes to rebrand information literacy for the AI age. Coins cute acronyms hoping they'll catch on like CRAAP. Publishes "AI literacy frameworks” in journals that boil down to "think critically" and "verify sources"—things Information Literacy has always done, based on nothing but “I thought it obviously makes sense”. Same talk given seventeen times: "AI and Information Literacy: Challenges and Opportunities."

(Often evolves into Group 8 once they realise people will pay. Frequently co-occurs with Group 7—needs to try tools to have something to teach.)

Group 7: The Tool Magpie

"Have you tried Perplexity? Wait, no—Gemini's better now."

Has tried every AI tool released in the past eighteen months. Switched from ChatGPT to Claude to Gemini to Perplexity to Copilot and back again—this week. Pays for seven subscriptions, uses none deeply. Confuses tool-hopping with expertise. Posts breathless first-impression threads within hours of each release; has never used anything long enough to find its actual limitations. Evaluates tools by vibes and UI polish, not outputs. Cannot tell you which tool is best for a specific task because they've never stuck with one long enough to know. Knows everything about AI in general, nothing about any tool in particular.

(The shallow-enthusiast wing. Natural feeder pool for Groups 6 and 8—surface knowledge monetises easily.)

Group 8: The AI Grifter/Consultant

"Join my masterclass: AI for Librarians—early bird pricing ends Friday!"

Runs paid workshops, webinars, and "masterclasses" on AI for desperate librarians who want to “learn AI”—at $200 a seat. Content is surface-level: how to write a prompt, what ChatGPT is, maybe a live demo that could go wrong. Repurposes freely available material behind a paywall. LinkedIn posts optimised for engagement: "🚀 5 AI prompts every librarian needs to know!" Has a Canva-designed PDF lead magnet. May offer "certification" of dubious value. Credentials for teaching AI: used ChatGPT; spotted anxious librarians willing to pay. The hustle is real, even if the expertise isn't.

(The final evolution of Groups 6 and 7. May have started as any type—including Group 1—before sensing the market.)

Group 9: The Eternal AI Apologist

"This is the worst AI will ever be."

Every limitation is temporary. Every failure is a version number away from resolution. Hallucinations? Fixed in the next release. Bias? They're working on it. Can't do maths? Have you tried the new model?

Treats OpenAI's roadmap as prophecy. Believes "reasoning" and "agents" will solve problems that are actually fundamental to the technology, not bugs to be patched. Has been mass-adopting an "early version" for three years now.

Will dismiss any criticism with "that's old news"—even if the paper came out last month. Cannot distinguish between genuine improvement and better benchmark gaming. The goalpost doesn't move; it simply hasn't arrived yet.

When the technology fails, it's never the technology. Everything is a skill issue. You prompted it wrong. You didn't use chain-of-thought. You should have tried the API instead of the consumer product. The tool is fine; you're just holding it incorrectly.

Favourite rebuttals include "you sound like people who doubted the internet, web search engines, and Wikipedia" and “you are just a luddite”

Never explains why, if improvement is so inevitable, we shouldn't just wait for the version that actually works.

(The mirror image of Group 5. Equally unfalsifiable. Will accuse critics of being Groups 1–4 without recognising they've formed their own church.)

Group 10: The “Visionary” Director

"We're leveraging AI to transform the future of libraries."

Has given 19 keynotes on "AI in libraries" but whose own experience with the technology is using it to draft an email. Peppers presentations with buzzwords—"generative," "agentic," "multimodal," "AI literacy," "cognitive offloading"—absorbed osmotically from vendor webinars and conferences. Cannot explain what any of them mean but knows they sound forward-thinking.

The organisation's entire AI strategy: a Microsoft Copilot enterprise licence purchased eleven months ago. Usage stats remain in the dumps. Somewhere, a single staff member used it once to draft a meeting agenda.

May hire a consultant (group 6 or 8) to provide “training” once and consider the job done.

Conference bio describes a "pioneering approach to AI integration." Integration consists of adding "AI" to the strategic plan and forming a working group that has met twice. The working group awaits direction.

Has no vision or concrete direction but knows innovation is happening because a committee was formed to discuss it.

Asks frontline staff what AI tools they're using, then presents their answers at the next conference as institutional achievements. Credits "our team" while being unable to name anyone on it.

(Group 6 or 8s can evolve into this. Alternatively, some provides political cover for doing nothing while appearing to do everything.

Bonus: The Professional Sceptic blogger

"Interesting, but the methodology is opaque."

Has tested every AI search tool and found them all wanting. Blogs about vendor overreach and methodological opacity—valid critiques, but you wonder if anything would ever be good enough. Tries to actually explain the technical terms based on things he just looked up while blogging and naturally messes it up.

Considers him/herself as the perfect balance of open-minded and skepticism,

(Evolved Group 7. May overlap with Group 5, but with better receipts. Often enemy of all other groups; Worst of the worst)

This is all tongue in cheek. Any resemblance to a real individual is purely coincidental (except maybe Group 9).

I'm clearly the last plus big parts of 7, 8 (I hold paid workshops!) and 9.