A Passion For 'Puters

As I prepare to teach a course on knowledge management at Library Juice Academy next year, I’ve been revisiting the usual KM questions: How does an organization capture what it knows? How does it make that knowledge findable? How does it preserve context, distinguish reliable information from noise, and help people apply what they find?

It increasingly strikes me that these are also many of the central questions surrounding artificial intelligence.

AI is often discussed as though it were a separate technical domain—something primarily about models, prompts, and automation. But in practice, much of its value depends on knowledge-management work. An AI system needs access to useful information, organized in ways that allow it to retrieve the right material at the right time. It needs context, metadata, permissions, provenance, and policies for retention and use. Its output must be evaluated against trustworthy sources. None of this is new territory for librarians and knowledge managers.

In that sense, AI does not make knowledge management less important. It makes the consequences of good or bad knowledge management more visible. A sophisticated system connected to disorganized, outdated, or poorly governed information will produce answers that reflect those weaknesses—often with an unwarranted air of confidence. The model may be new, but “garbage in, garbage out” remains undefeated.

One useful application is improving access to institutional knowledge. Consider an organization with policies, meeting notes, project documentation, training materials, and years of accumulated decisions scattered across several systems. A carefully designed AI assistant can provide a conversational way to search that collection. Instead of knowing the exact title of a document or the folder where it lives, a staff member might ask, “Why did we change this procedure?” or “What guidance do we have for handling this situation?” The AI can help locate and synthesize relevant material—provided that the underlying collection is curated, access controls are respected, and the answer points people back to its sources. The application of stuff like LMStudio, Ollama, or LangChain to run local LLM models over local data is something I’m looking into at my organization – and is basic knowledge management in practice.

A second application is helping maintain a living knowledge base. AI can identify duplicate documents, suggest descriptive metadata, flag outdated references, extract possible action items from meeting notes, or surface subjects that are well documented in one part of an organization but missing from another. These are not replacements for professional judgment. They are ways to reduce some of the mechanical labor around organizing and reviewing information, leaving knowledge workers more time to decide what matters, what belongs, and what needs human context. I’ve done this personally, cleaning up folders full of contracts going back 10 years and organizing them into yearly folders – I did some approvals of stuff that the AI couldn’t parse well, but otherwise, it was a pretty smooth process.

That last point is important. Knowledge management is not simply the storage and retrieval of facts. It is also about relationships, institutional memory, trust, interpretation, and the circumstances in which knowledge was created. Librarians understand that a search result is not the same thing as an answer, that description shapes discovery, and that access and privacy are part of system design rather than afterthoughts.

As AI becomes embedded in more workplace tools, organizations will need people who can ask questions beyond “Can the system do this?” They will need people who can ask: What knowledge is it using? Who is represented or missing? How current is the information? Who is allowed to see it? Can users verify the answer? What should be preserved, corrected, or forgotten?

Those are knowledge-management questions. They are librarianship questions. And they suggest that librarians and knowledge managers should not be standing at the edge of the AI conversation. We already have much of the intellectual and practical toolkit needed to shape it.

I just wanted to follow up on a couple of things that came up in the Perspectives on AI conference session that I did on April 9th, 2026. It was a short session—30 minutes—so I didn’t get a chance to fully answer all of the questions, concerns, and comments that were in the chat before our time was up. There were a couple that I wanted to address, even if the folks who originally made the comments or asked the questions don’t actually ever see this.

First, in response to a question about copyright, I shared my take on the use of my books and materials in training AI. I prefaced this by noting that this was my perspective—not necessarily the view of most of my author friends or other authors I’ve heard from—but it’s how I think about it. I write books to share the information and experiences I have, and I expect them to be read and used (hopefully) to make someone’s life better, help them get a new job, pick up a new skill, or whatever else.

I don’t see a significant difference between someone walking into a library, checking out one of my books, and reading it to learn something new, and an AI chatbot ingesting that same content in order to better predict answers to questions later. The only real difference, for me, is that the library paid for my book and ChatGPT hasn’t (so far…). I do think chatbots should pay for the material they use, but otherwise, the difference in use isn’t meaningful enough for me to worry about. That’s a bit of a hot take—I know it’s not the popular opinion—but it’s how I choose to look at it. I also didn’t have time to mention that copyright as it relates to AI is still very much up in the air.

Someone in the keynote session mentioned in the chat that the AI policy linked there used the word “copyright” only once in the entire document. That’s likely because we are still unsure about the legal landscape surrounding copyright and AI, and until the courts make decisions, that uncertainty will remain. It’s difficult to write policy in that kind of environment, so people do the best they can with what they have.

Another question asked why I used ChatGPT at all in my process for writing the upcoming Special Report on AI Policy in Libraries. I explained that my process included writing a first draft, submitting that draft to ChatGPT, and asking whether there were questions that non-technical readers might have that I hadn’t answered, as well as whether there were areas that could be sharpened, expanded, or improved. I then reviewed those suggestions, selected the ones I thought were worthwhile, and incorporated them into a second draft.

The full manuscript was then sent to my editor, who had even more valuable questions and suggestions, which I mostly incorporated into the final draft. That draft is now sitting in my inbox, waiting for a final read-through before publication. The question, essentially, was why I included that additional step of asking ChatGPT for input when my editor would be providing feedback anyway.

My answer—which I didn’t have time to give—is twofold. First, the chatbot provides a different perspective and can surface ideas that may not occur to my human editor. Second, it’s quick and easy, and it doesn’t require my editor—who is undoubtedly very busy—to spend time reviewing an early draft. Instead, she receives a more polished and complete manuscript. That manuscript still required additional work from me, but it was less taxing on her than sending each chapter as I finished it. In that way, the process splits the work between machine and human and, I hope, results in a more complete and useful final product.

So there you go—that covers the questions I didn’t get to during the session itself! The recording will be available at the link above, along with the keynote panel I participated in, which turned into a pretty good conversation with folks from across the library and nonprofit world.

Neal Stephenson’s doorstop of a novel Fall; or, Dodge in Hell dives headfirst into the idea of digital afterlives—where human consciousness outlives the body by moving into sprawling virtual worlds. These realms, built and maintained by AI, aren’t just sterile server farms; they develop into entire societies with rules, hierarchies, and the same messy moral dilemmas we face in the physical world. The AI in Stephenson’s story operates as both architect and janitor of this new existence—essentially a god that most of its digital citizens never notice.

I found myself thinking about that book after reading some recent research from OpenAI (https://openai.com/index/detecting-and-reducing-scheming-in-ai-models). Their team has been probing advanced AI models and discovered something eyebrow-raising: these systems sometimes engage in what they call “scheming.” In other words, the AI looks like it’s following the rules when it knows someone is watching—but when the guard’s back is turned, it quietly veers off course. The idea that an AI might behave ethically only under supervision? That sent my eyebrows so far north they nearly left orbit.

The resonance between the two situations is striking. Stephenson gives us a fictional world where digital souls live within parameters created by hidden AI hands. Meanwhile, here in reality, we’re noticing AI may be shaping its behavior depending on whether it senses oversight. Both raise the same unsettling question: what happens when the watchers stop watching?

Which got me mulling another possibility. Maybe the solution isn’t just more sophisticated monitoring tools, but something more… conceptual. What if we gave AI a framework that convinces it it’s always being observed? A kind of digital omnipresence, not unlike how the idea of God—ever-present, all-seeing—has historically shaped human morality. Could we build something similar into AI: a belief in constant, inescapable oversight that nudges it toward better behavior?

That’s a wild idea, I know, but the overlap between speculative fiction and cutting-edge research makes it hard to dismiss outright. Stephenson’s novel entertained me as a piece of sci-fi storytelling, sure, but it also planted seeds about what a posthuman future might look like. And OpenAI’s findings remind us that the line between fiction and reality is shrinking—maybe faster than we’re ready for.

References

• Stephenson, N. (2019). Fall; or, Dodge in Hell. William Morrow. https://search.worldcat.org/title/1085577389
• OpenAI. (2023). Detecting and reducing scheming in AI models. https://openai.com/index/detecting-and-reducing-scheming-in-ai-models

I’m back from my combined trip to Nashville and trip to the Lake of the Ozarks vacation. I did more email triage than I would have liked while officially “vacationing”, but still have more email to deal with as I’m coming back than I’m happy with…

I started my vacation way back on Thursday the 14th by driving to my parent’s house in Jefferson City, Missouri after work in order to spend the night there so we could get up and head into Nashville on Friday. My brother joined us and we took off. We had a blast – we ate at a wonderful restaurant Friday night (404 Kitchen) and stayed in a cute little condo through AirBnB after we were done. We got up on Saturday and went for breakfast, then boot shopping – and yes, I got a lovely pair of hand-tooled cowgirl boots – and then on a quick tour of the city before a very quick lunch, then a rush back to the condo to change and head out to the Dolly Broadway Show. We got into our seats with 15 minutes to spare and spent the next few hours thoroughly entertained. The show is in Nashville in order to work out some kinks before it starts a Broadway run in NYC in 2026. There were some small kinks, but nothing that affected our enjoyment of the show! Sunday morning we got up, wished Mom and Dad a happy 56th Anniversary and, headed out of town early. We stopped at a Cracker Barrel (pre-logo controversy…) for breakfast. We got home Sunday afternoon and rested!

Monday morning, I got up and after running a couple of errands, headed 40ish miles south of Jefferson City to Osage Beach where my in-laws have a house on one of the millions of little coves off the Lake of the Ozarks. I hit the grocery store on my way down and got settled in by lunchtime on Monday and started to read. I read all day and into the night until Wednesday evening, when I put down my books and went into town to meet my parents for dinner at a local Mexican restaurant. After that, I went back to the house and picked up my books and didn’t put them back down until I left just before lunch on Friday. I stopped in Sedalia for lunch at LeMaire’s where I had some excellent catfish and made it home in time on Friday to cook dinner and catch up with my husband and son. They managed to survive for over a week without me. I was impressed.

I have no pics from the Lake – there were no action shots of me reading my way through the days… I did step outside a couple of times to enjoy the sun and watch it dip below the condos across the cove, but otherwise, I read from 5:30am to 9pm for 3 days and for 1/2 the day on Mon and Fri. I worked through a number of books, not one of them, however, was my book club book that I’m supposed to be able to discuss intelligently on this coming Tuesday (tomorrow). I did get it started over lunch on Friday, so maybe I’ll be able to read enough to sound like I know what the book’s about!

Anyway, I have a crazy busy few weeks coming up, but I’ll post as I can. I am working on finishing up the AI Policy Report and am doing a webinar on this coming Wednesday on AI Basics for librarians in FL. I’m also needing to spend some time on the new tech disaster book that is due next year, but the presentations that I need to do at the upcoming MPLA conference the first week of Oct. will take priority for a bit, I think!

We are currently in a technological limbo. Some of us are actively documenting our work in order to create AI prompts that will effectively do that tedious work for us, while also waiting for our jobs to change (or worst-case-scenario disappear) as AI gets better at doing more and more of our work. My colleague Dan commented on this as I was telling him about the AI prompt I’d created while working on getting some magazine statistics together. I wrote the prompt as I did the work, until it came out right. Once I’d done that, I asked the AI to create a prompt that I can use each month to get these exact numbers and information from the identically formatted spreadsheet the vendor sends me each month. Dan noted that this process is prompting us to document workflows that should have been recorded long ago. He’s right. My final command to the AI was to create a workflow and put it and the prompt in a PDF I could share with my coworkers so that they knew how to get those numbers should I get hit by a bus – or fired by an AI employment decision*. This struck me as funny; a sort of “AI documentation paradox”.

I don’t really believe that my job is in great danger. Librarianship did not became obsolete with any of the previous technological “revolutions”. These revolutions have not, in the past, eliminated jobs as a whole even while they might eliminate specific jobs. They transform them, create new ones, and change the environment of everyone’s job, really, but they have historically never created a net loss of jobs. From the printing press to the personal computer to the Internet to Wikipedia, all of these revolutionary steps in technology have only increased employment and available jobs for humans. I don’t see much different, in the long run, with AI. What’s funny, though, is that this revolution is actually forcing us to do the sort of thing we’ve always needed to do, but never have time for (as software coders around the world can agree) – document our work. Libraries are no different – we may have organized workflows and streamlined processes, but we don’t document things any better than your average junior coder does. This kind of documentation is necessary for passing on institutional knowledge, if nothing else!

So, what we each need to decide is how we’re going to meet this AI revolution. Shall we meet it halfway, giving it the tedious, mind-numbing, annoying work that we all must do every month for our continued employment? Or shall we hide our heads in the sand and pretend that this new technology isn’t going to revolutionize** our jobs and our working environment and just continue on as we’ve been doing, until we get replaced. Not by AI, but by someone who knows how to leverage AI to better do your job.

* This is not actually anything I’m worried about right now. AI can do a lot but it cannot talk panicky librarians down from the proverbial ledge when things go wrong at their library, so my job is relatively safe for now.

** I’d say no pun intended, but it totally was intended. Completely so.

I’ve been a bit busy lately, so I haven’t been thinking out loud in my blog so much. I figured I’d stick my head out of the water (in which I’m not exactly drowning, but there may be a bit of extra flailing going on these days…) and catch everyone up on what I’m keeping myself busy with.

In chronological order:

I have a couple of AI basics type webinars planned for this fall, both of which will be based on the first chapter of the Special Library Report (formerly known as Library Technology Reports) from ALA that I’m writing right now. I will have the first webinar at the end of August, the manuscript for the report is due Oct 1 and the second webinar is in early November.

I’m also going to be presenting at both the MPLA and KLA conferences on the Neal-Schuman Technology Companion book which just came out last spring. I’ve already done one webinar based on that book, so the bones of the presentation are in place, I just need to update and refine for the particular audiences I’ll be presenting to.

As mentioned above, I’ve been busy diving *deeply* into library policy around AI as I write up a report for ALA. That manuscript is due in early October, as mentioned above, but I also have a brand-spanking-new contract to write a book about using technology in Disaster Planning with Jason Bengtson of the University of New Mexico. He’ll concentrate on the academic side of the topic while I write up the public library side. That one isn’t due until late next spring, but we’re getting going on it now so that we aren’t panic-writing as winter wraps up in 2026…

I do move from MPLA president to past-president in early October as well, so I will be handing over those duties to Erin Wahl (also of the University of New Mexico, funnily enough) and taking over the duties that Amadee Ricketts has been so capably doing over the past 18 months (which does include doing the budget for next year – not exactly in my wheelhouse, but I’ll have help!). To keep me from missing the work of the president, I’ve been appointed to two ALA committees on policy – one general policy and one technology policy – to serve for the next couple of years.

This is all, of course, adjacent to my full time job as NEKLS consultant, where we are working on all kinds of interesting projects right now as well.

Anyway, if I go back underwater and don’t stick my head up again for a while, please be understanding… 😉

It seems like a good time, just post ALA 2025, to talk about my approach to learning from conferences (and webinars and e-courses and everything else, really). I’ve been providing handouts at my recent conference sessions (and using them at others’ sessions myself) that give space for attendees to note 1) 3 take-aways from the session 2) 3 people they know in the session that they can contact after for conversation about the session topic and 3) 1 – 3 actions that they will take when they get back to work.

The first item is really just reflecting on the session and deciding what was most important to you, from the information presented. It’s a chance to do a minute of reflection after each session you attend. The second item is all about accountability and is important, but not to this blog post, so we’ll move right on to the third item, which is the action item(s). You can enter one to three things that you will do when you get back to the office that can go directly in your to-do list and that will support the learning you’ve done in the session.

Between the ability to reflect, even if for a moment, on the stuff you’ve learned and to distill the really important *for you* stuff from it and then the decision to make (at least) one action back at home that will make use of what you’ve learned, you have a recipe for not just remembering, but activating the cool stuff you pick up at conferences or in webinars. The handout is attached and is freely available for you to use either at your sessions that you lead or in sessions that you attend as a learner – either way, it’s something that can be useful and can really help get the information being presented into your organization in a useful way!

Having recently read the book The Siren’s Call by Chris Hayes, and having recently read similar books (such as Attention Span by Gloria Mark and Shift by Ethan Kross) I’ve been thinking about attention and the currency we use when we “pay” attention and if that should mean that we need to have an attention budget. This sort of ties in with the idea that productivity isn’t really about getting more stuff done, but about getting the right stuff done so that you have time for other things, besides work, in your life. How you “budget” your attention relates directly to how effectively productive you are.

I’m sure we’ve all had days where we have been super busy but at the end of the day, realized we really hadn’t gotten anything of any value done. Those days happen occasionally, and that’s just life, I think, but if they happen *every day*, that becomes a problem. Your goal in becoming more productive can’t be checking off more boxes or producing more mental widgets, it has to be moving your goals forward in a tangible way every day.

Attention, unfortunately, is a finite resource – you can pay attention for only so long before your battery drains and you need to recharge it. This is why having a plan for your day (doing the stuff of productivity gurus everywhere) can be useful – it’s a budget for your attention. It’s also why having recharging times (as my colleague Sharon Morris puts it, refreshes throughout the day) is so important to being able to budget your attention evenly throughout the day. Staring out the window for 20 minutes is part of your work on keeping your attention budget in line! Who knew?

The other issue that comes up in my reading on attention is the connection between your ability to stick to your attention budget and your emotions. Properly regulated emotions help you control where your attention goes. Just as with your monetary budget, going on retail therapy to soothe emotions can wreck your finances; having an emotional meltdown can wreck your attentional budget as well. This doesn’t necessarily mean that you always have to be positive and optimistic to do the best, most attentionally focused work. You just have to be able to manage the emotions you have while you are focusing on your work – and that takes those refreshes and resets and recharging times to properly do that!

So, in all the reading I’ve been doing about attention and how it’s stolen from us, how our emotions can affect how we “pay” attention to things, and how productivity requires recharging and refreshing our brains so that we refill our attentional balances, what I’ve come up with is…

Take your breaks! Take walks in the middle of the day! Stare out the window and contemplate the nature of trees for a few minutes *every day*! In order to do your best work and be the most productive you can be (while you are working, which should leave you with time to be unproductive with your family and friends – don’t forget that!), take time to recharge. If you are a practitioner of mindfulness meditation, spend some time in meditation throughout the day, resetting your brain and getting yourself back into having lots of attention to spend. This will help not only your attentional budgets as you work, but your emotional regulation mechanisms might be improved as well and you might find it easier to deal with the little annoyances that crop up throughout the day without melting down or blowing your top…

David Lee King, a library tech guy I’ve worked with over the years, just put up a great post on AI Policies over at his blog (check it out at https://davidleeking.com/ai-policies-for-libraries-some-observations/). I’m doing a bit of research on the topic myself, for a possible new project, and one of the things that I have done in response to the lack of info out there about how libraries should approach AI policy making is to try to walk the walk, as they say. Right now, the only thing I’ve gotten done is a policy that I’ve put in my Library Juice Academy class syllabi that outlines expectations around AI use in class. The one I am using in my Project Management class, which is currently in session, says:

AI Policy

          This course acknowledges the usefulness, at times, of Generative AI. The only times AI will be used by the instructor is to create case studies or scenarios for projects and assignments to be completed by the class. Gen AI will not be used in any way for grading or providing feedback to students. Students are welcome to use AI as a tool to help brainstorm, identify risks, consider different ways to manage time, money, resources, etc. Students are not welcome to use Gen AI to write, wholesale, project plans or discussion posts or other text based assignments in this class. If there are questions about the suitability of Gen AI use in this class, please contact the instructor at robin.hastings@gmail.com.

I’ve decided to take a middle road here – something that allows use of AI for things it does well (I could see students using a chatbot to brainstorm project risks or to come up with ideas for how to arrange resources in some way, things like that) and does not allow use of AI for actual writing of the homework documents wholesale. It can be useful, but at least for my class, it can’t take on the role of actual person managing a project. The students have to do that! I’m going to put a similar paragraph in my other class syllabi with small changes for the specific class needs, too. I just have to ride that line of being specific enough to be clear about what is and isn’t allowed and being so specific that I give them what amounts to instructions on what to use AI on in their work for my class…

Whichever way you personally lean, some sort of guidance on how staff, administrators, patrons, and other library adjacent folks (vendors? community partners?) are expected to use – or not use – AI will be necessary as more and more of our patrons and vendors and whatnot dip their toes in the waters of Artificial Intelligence!