Infophilia: A Positive Psychology of Information | May 30, 2026 | Vol. 4, Issue 28

Biophilia |Library and Information Science | Wellbeing

✨ Welcome to Infophilia, a weekly serial and a living lab exploring how our love of information and connection can help us thrive, individually and collectively. This is avant-garde research in how we engage with knowledge, meaning, and each other. If this is your first time here, I’m glad to have you. To learn more, browse the links above, the Archive or About. Or simply discover your infostyle by taking an anonymous, private, 10 question assessment.

📌 Access & Attribution: This serial offers free previews always and occasional open access essays to keep scholarship accessible. Students and those facing financial barriers can request complimentary access. If you find value here, please cite the original work and consider supporting the public scholarship through subscription. Proper attribution sustains this work and models healthy information engagement.

Cite as: Coleman, A. S. (2026). How an AI response is generated: An Information Story. Infophilia, A Positive Psychology of Information 4(28).

We continue this week with a new Information Story — systems-aware and poetic rather than apocalyptic. But first, some news.

Several major library groups — COSLA (Chief Officers of State Library Agencies), the Canadian Urban Libraries Council, the Association for Rural and Small Libraries, and the Urban Libraries Council — have joined the Public Library Association to fight publishers on restrictive ebook pricing. Libraries pay about $55 for a two-year license on a title that costs a consumer $13 to own outright, and many now spend more than half their collections budgets on time-limited digital content that vanishes when the license expires. Reader wait times have surged. Several U.S. states and Canadian provinces are turning to consumer-protection laws. Who captures the gains, again.

Andrej Karpathy — the influential AI researcher and teacher, one of the eleven who co-founded OpenAI and led AI at Tesla, and the man who coined "vibe coding" (which we've discussed here twice) — joined Anthropic on May 19th, reporting to the head of pre-training. His assignment: build a team using Claude to accelerate the training of Claude. The industry term for where this leads is "recursive self-improvement" (AI improving AI). The loop tightens: human language trained the machine; now the machine helps train the next machine. Our words, one more layer removed.

The Stanford 2026 AI Index released last month is the field’s most comprehensive annual benchmark. And now we have the May 2026 AI User Metrics Report from Edison Research at SSRS which tracks AI awareness and usage across the U.S. adult population twice a month. It confirms what most of us sense: ChatGPT and Gemini dominate, while Claude’s share has more than doubled. AI users still skew toward higher-income households. The tools are not yet evenly distributed, and neither are the gains.

Kyiv hosted the XIII International Book Arsenal Festival, highlighting translation and culture. Translation has important implications for intellectual freedom: many words cannot be translated. Language creates reality.

“Send arXiv AI-generated slop, get a year-long vacation from submissions.” The physics and astronomy preprint server announced that inappropriate AI-produced content will earn a one-year ban and a permanent requirement that future submissions undergo peer review first.

A reader forwarded a remarkable notice from IGI Global’s VP of Publishing Ethics. The publisher has been auditing social media and finding chapter titles with open “authorship spots” — chapters already written, often by AI, inviting individuals to pay a fee to attach their name. Authorship for hire. Chapters for sale. The publisher is now locking author accounts at submission and refusing declarations of authorship as proof. Yes, read that twice: a declaration that you wrote your own work is no longer accepted as evidence that you did. That is the trust environment AI has produced.

The same reader sent us a new preprint — broad-based, not yet peer-reviewed — that found that human-written essays offered up to eight times more new ideas than those produced by AI. The machine gave an illusion of creativity through novel terminology, but the substance was mundane. Which is exactly what this week’s story is about: AI gives the illusion of novelty while being, fundamentally, infrastructure speaking.

We close with an item that says everything: South Africa’s national AI policy cited studies that don’t exist. AI wrote the citations. A government, proud of its forward-looking policy — until it discovered the policy had hallucinated its own evidence.

Which raises the question this week’s story asks: what exactly is an AI response?

How an AI Response Is Generated

An Information Story

Read more

Infophilia: A Positive Psychology of Information | May 23, 2026 | Vol. 4, Issue 27

Biophilia |Library and Information Science | Wellbeing

✨ Welcome to Infophilia, a weekly serial and a living lab exploring how our love of information and connection can help us thrive, individually and collectively. This is avant-garde research in how we engage with knowledge, meaning, and each other. If this is your first time here, I’m glad to have you. To learn more, browse the links above, the Archive or About. Or simply discover your infostyle by taking an anonymous, private, 10 question assessment.

📌 Access & Attribution: This serial offers free previews always and occasional open access essays to keep scholarship accessible. Students and those facing financial barriers can request complimentary access. If you find value here, please cite the original work and consider supporting the public scholarship through subscription. Proper attribution sustains this work and models healthy information engagement.

Cite as: Coleman, A. S. (2026). Health literacy: An Information Story. Infophilia, A Positive Psychology of Information 4(27).

This week I'm launching Information Stories, a literary format I've been developing alongside the book, Infophilia Unbound. Each story takes something you encounter in everyday life and walks you through the systems behind it: the infrastructure, the incentives, the human choices, and the possibilities. They’re short enough to read in one sitting. They end with agency and imagination. And they’re designed to help you see the world you’re already living inside more clearly. The first is about health literacy and it is for anyone who has ever left a doctor's office more confused than when they arrived.

Information Stories are also, not coincidentally, a revenue format for AI agentic access — yes, we know we're ahead of the market. But are we? At Google I/O 2026 this week, Sundar Pichai announced what the company calls the biggest change to its search box in 25 years. Search is moving to AI Mode — conversational, agentic, multi-modal, and designed to keep users inside Google. Zero-click searches already account for roughly 60% of all queries. Publishers, bloggers, and independent creators who depend on Google referrals are sounding the alarm.

So why did I choose health literacy?

My dissertation was a hypermedia journal — text with processing tools built in — designed for health information workers in continuing education. Health literacy has never left me since then. And some of my readers here are health professionals who’re keenly following the new National Institutes of Health-Wide Strategic Plan for FY 2027-2031. Well, this week, the American Historical Association submitted a comment to the Federal Register supporting the National Library of Medicine, expressing concern about recent reorganizations including the elimination of its History of Medicine Division. The AHA argues this signals a departure from long-standing commitments to preserving the nation's biomedical history and raises serious questions about future access to critical materials. Health literacy depends on institutional memory.

Information Stories launches between two significant dates: May 1, which independent creatives mark as Fall of Freedom Day, and Memorial Day on the 25th.

Appropriate timing.

The question of who captures the gains from AI productivity is becoming impossible to ignore. In South Korea, Samsung's union at the world's largest memory chip company reached a landmark AI bonus deal tied to operating profits, while a top policy aide proposed a "citizen's dividend" — a share of AI boom profits distributed more broadly. In the UK, reporter Graham Lovelace (Charting the Gen AI Substack) has covered the government's consideration of a text-and-data-mining exception that triggered major backlash from creators, publishers, and campaigners — and a government retreat. Both are versions of the same question: who captures the gains from AI systems built on copyrighted work and human labor? If you watched open access get commercialized from the inside, you know exactly how this story goes.

I also want to be completely candid. Information Stories are only for paid subscribers. Independent scholars have to earn a living and Infophilia from inception has always been a paid subscription. Some of the news from just the first 20 days of May has helped reinforce that decision.

1. The Economist is one of my favorite reads but I read it through PressReader (a database or app that bundles popular newspapers from around the world, and available remotely through my library). And now The Economist is describing ideas from largely abandoned hypermedia research as the future of the internet. This week, they announced that they are preparing for a two‑track internet: one for humans and one for AI agents. Information Stories are the perfect format for both groups and this means that independent scholars like me, writing for a living, need to protect ourselves. Readers who value creatives must once again choose to support them and unlearn and relearn what the Internet taught us to forget: you pay for what you get.

2. And it’s already happening in our own backyard. On May 19th, EBSCO Information Services announced a partnership with Perplexity — the AI answer engine and maker of Perplexity Computer, a multi-step research agent — bringing EBSCOhost research databases into Perplexity's Premium Sources. Perplexity’s Publisher Program offers an 80/20 revenue split to participating publishers. The pattern is familiar: revenue flows first to institutions and aggregators. Independent scholars and creators are still waiting. Reliably. As always. Open access taught us that lesson too.

3. You’ve probably already heard that commencement speakers who mentioned AI were met with boos at institutions such as the University of Arizona — where I was a full-time faculty and developed dLIST, the first open access disciplinary repository for LIS, still quietly used and largely unknown. Well, Ann Kirschner writing in Forbes put it well: “Boo it and Use it: What Gen Z Gets Right About AI.” The problem is not Gen Z rejecting AI so much (they are using AI more than any other group) as what people with power reliably do when “institutional realities” are invoked. Kirschner’s example is instructive. She points to neighboring Arizona State University Professor of Practice — the Black Eyed Peas frontman, yes, that will.i.am — who is teaching a course “The Agentic Self” and has this blunt take: “You could use AI or AI could use you.” And she adds her own observation that Gen Z is already “using the tools, doubting the hype, making their own calculations… The booing at commencement wasn’t only about AI. It was also [about] being lectured by people who presented a formulaic perspective on a complicated future that the class of 2026 is already navigating.”

4. ShinyHunters. TeamPCP. A shiny hunter used to be a Pokemon player's obsessive quest for a rare gleaming creature (that beloved game whose culture is now global and which we playfully co-opted to describe librarians earlier this year). Now, it’s the name of a malware group that claimed responsibility for the Instructure breach that led to Canvas breaking down during exam week in major universities. Then, TeamPCP that broke into GitHub (I call it HitHub now) “reckons it now has access to GitHub's source code.” Even the house of code wasn't safe from the hunters.

That’s why Information Stories are not open access.

An Invitation

If any of this sounds like gobbledygook join us for the next Infophilia Book Club / Office Hours on Wednesday, 17th June, 12 noon Pacific / 3 pm Eastern. We’ll be discussing Emily Knox’s Book banning in 21st century America (2nd edition) published just earlier this year by Bloomsbury. Knox is one of the most important voices on intellectual freedom working today and this second edition couldn't be more timely, beginning as it does with her testimony in DC and insights into “Challengers.” We’re also happy to answer questions and discuss these essays. Just contact me for the Zoom link. All readers are welcome — free, paid, patrons — whether or not you’re able to read the book.

Spring has come to a close. Summer is bringing something new: Information Stories.

Enjoy your first Information Story and have a lovely Memorial Day Holiday,

Anita

Read more