The number of colleagues across disciplines who are speed running and openly delighted at describing how they are offloading tasks that are the core of what we as faculty are privileged to do — creating new knowledge, writing up our findings, sharing our knowledge. There’s always been the “oh and I have to teach” aspects of it for many programs that are training researchers but now there’s also this seeming desire to not do their own research too? It’s unclear what they anticipate their job is to be.

Anyway, an internal colleague asked for some thoughts related to what I was thinking about as we face cases where junior faculty are navigating the weird pressures of “you must AI OR DIE” versus “so you’re offloading your critical thinking to an ethically corrupt and ecologically disastrous auto-complete calculator.” Here’s a slightly edited version of what I sent her — with a recognition that this is not a comprehensive document but questions I am noodling on…

1. We need to define “AI” and stop lumping everything under a single umbrella term. Are we talking about use of spell check and grammar check that catches the wrong their/they’re/there; testing novel machine learning or algorithmic tools that are working across datasets too large to identify correlations; automating some data clean up and standardization; or generating entire literature reviews and inappropriate rat visualizations? There are significant differences in perceived usage varying from “My campus version of Office 365 has spell check and won’t let me turn off autocomplete suggestions” versus “Here is this manuscript, can you tell that I used an extruded text from Co-Pilot for the literature review and I had it “do” the analysis?” The Artificial Intelligence Disclosure (AID) Framework by Karl Weaver may be useful here.
   
2. Faculty will need to put in more effort with their manuscripts if they plan to use an LLM.  There is no reason that mentors, peers, peer reviewers, editors, or other colleagues should be asked to put in time and effort reviewing and providing feedback on extruded text manuscripts that you couldn’t be bothered to write, particularly when those manuscripts are half-baked. I’m seeing this with students and it’s creeping into peer review. You are not too busy or important to meaningfully engage in drafting your own manuscript.
   
3. The often very obvious use of LLMs to draft literature reviews suggests a lack of curiosity and engagement with the discipline and your colleagues – if you aren’t reading and engaging with your peers, why are you producing materials that you expect them to read? How can you be sure this research actually meets a gap in the literature? How are you critically engaging and creating a story to lead to your question rather than writing slightly shorter sentences than what is in the abstract? It suggests a misunderstanding of the point of a literature review that is not flattering.
   
4. Use of LLM tools is likely to reinforce citation biases and has a strong potential to enhance erasure of women, faculty of color, and other minoritized groups in the literature across disciplines. The tools are trained on what is available to them and we know there are significant problems there. The answer cannot be to steal more of our currently copyrighted materials to train a tool that is then sold back to us. The LLM companies are also ingesting retracted papers without capturing the metadata of retraction, which will cause problems where someone aggressively asserts something that has been refuted. Systematic reviews and meta-analyses will get worse and less trust-worthy. This is on top of the disappointing behavior of researchers not bothering to verify that a citation even exists before submitting it to a journal.
   
5. The role of tenure and scholarship at our institutions is to create new knowledge. I was given tenure not only on what I had accomplished to date –which absolutely mattered — but because the university believed I would continue generating new knowledge and work that impacted my discipline, the community, and the world. We’re already hearing about how the voice of scholarship is shifting to sound like the popular LLMs, the ideas are becoming homogenized because of the reliance on these tools. If everyone is asking the generally same prompt — as evidenced when the students all had an LLM create a personalized near-identical apology for their professor — we are not gaining new knowledge and ideas or developing new voices.
   
6. LLM tools cannot be authors – as is in the COPE Position, and use of LLMs must be weighed against ICMJE or similar mechanisms to determine whether a faculty member meets the qualifications for authorship. With researchers choosing to copy and paste materials without even removing the prompts that were used or the final closing comments at the end of the LLM output — can we say those authors have meaningfully taken responsibility for the final text? “I prompted a computer and maybe did some Make It Sound Human editing” doesn’t strike me as exceeding the sweat-of-the-brow argument for copyright. (There is a separately question about intellectual property and who it resides with if someone is using one of the open LLM models but I will leave that to the lawyers for today).
   
7. The use of these tools has been shown to have higher risks of data falsification as the various algorithmic tools invent charts, fill in data that doesn’t exist, etc. Faculty will need to verify this consistently and document that added data hasn’t been introduced or fabricated. This is particularly critical as the tools are written to return a confident-sounding, even if entirely false or non-existent or changed result. We have some but inconsistent consequences for publishing false data and these may need to be increased as this is further damaging public trust in scholarship.
   
8. For junior faculty, reliance on LLMs has the potential to demonstrate a lack of independence as a scholar. Use of these tools needs to be justified to address how you are overcoming this and why institutions should invest in researchers rather than solely having one senior faculty member just prompt and extrude some new text. If anyone with a vaguely decent prompt can generate your entire research agenda, why should the institution invest in expensive humans or award tenure? The belief in this is already floating around and is likely to lead to further de-professionalization and adjunctification.
   
9. We need to provide documentation to external reviewers, potentially with annotations, about how and if generative tools were used.  I’ve already started asking for this from institutions in external reviews and I think this will become common. Here is another instance where you will be asking me to spend a significant amount of my time and expertise. If the candidate couldn’t be bothered in the preparation of their dossier to meaningfully invest in this process, why should I?
   
10. And while this is last and brief, it is no less important that as faculty, we need to consider the ethics of using tools that have enormous ecological impact with water, land, and power; requires extensive filtering labor from the global south; is showing significant risk for causing mental harm through sycophancy; and facilitates technoableism.

A Few Suggested Materials

• Weaver, K. (2024). The Artificial Intelligence Disclosure (AID) Framework: An Introduction. College & Research Libraries News, 85(10), 407. doi:https://doi.org/10.5860/crln.85.10.407
• Guest, O., Suarez, M., Müller, B., van Meerkerk, E., Oude Groote Beverborg, A., de Haan, R., Reyes Elizondo, A., Blokpoel, M., Scharfenberg, N., Kleinherenbrink, A., Camerino, I., Woensdregt, M., Monett, D., Brown, J., Avraamidou, L., Alenda-Demoutiez, J., Hermans, F., & van Rooij, I. (2025). Against the Uncritical Adoption of ‘AI’ Technologies in Academia. Zenodo. https://doi.org/10.5281/zenodo.17065099
• The Data Workers Inquiry Project from DAIR https://dair-institute.org/projects/data-workers-inquiry/
• Damian Williams “Virtual Power: Al, Ableism, Racism, and Fascism” at Philosophy, Disability, and Social Change
• Authorship and AI Tools.  COPE. https://publicationethics.org/guidance/cope-position/authorship-and-ai-tools
• Bender E and Dair A. The AI Con. https://thecon.ai/  [This book is where I’m pulling the phrase extruded text from]

Her statement to ACRL Insider is here: https://acrl.ala.org/acrlinsider/meet-the-candidates-yasmeen-shorish-2/

and to C&RL News here: https://crln.acrl.org/index.php/crlnews/article/view/27183/35017

I have known and worked with Yasmeen professionally for well over a decade. She was part of the team that developed the call for the Research Data Management RoadShow that I built all those years ago. I’ve watched her advocate and take action at the national level through ACRL, RDAP, SPARC, and other organizations. At the same time her work and leadership at James Madison has shone through, when a new colleague of hers eagerly shared how much she had learned from Yasmeen when we met at a conference.

Through it all, she has been committed to her values and ethics, pursuing big goals mixed with a healthy dose of pragmatism that I have always appreciated. She spoke ten years ago about the sustainability issues in data services; these days I see her bringing patron-centered care and thoughtful criticism to the hyped and overblown language surrounding AI and the other buzzwords du jour.

I believe Yasmeen will confidently meet the moment in the three year term ahead of her if elected. She has already been an outstanding and dedicated committee leader, mentor, and member within the organization and her continued leadership will be so valuable to our members. I hope you will consider voting for her.

(12/1 edit — also this list doesn’t include all of the work I’ve had to do to rework coursework for a graduate course to make it less tempting for students to go the route of LLM copy/paste, which I don’t allow and which I don’t want to have to immediately fail students for using)

1. Advocate for and education around lab/dept/college/institutional policies for the use of various AI tools for research and teaching. Students, staff, and faculty are looking for guidance on acceptable use, when not to use it, where the university finds this correct or not. Frankly, most of them are relieved when I say “it’s not an option on this assignment / project / etc” because it removes confusion.
2. Create and present lectures and seminars helping students and colleagues navigate the bias, the ethical issues, the expectations around usage. Typically I’m brought in as the critic or the grounded person countering the overpromises in the headlines. I have a slide deck, there are quite a few recommended readings. I have other colleagues working on “how it’s integrated in Scopus” and “useful integration of machine learning in healthcare.” There’s plenty of instruction air being sucked up by these tools.
3. Continuously read and maintain regularly updated reading lists around the ethics and issues of using these tools. I have an entire book bibliography on this topic (help yourself, tell me what books I’m missing) and that doesn’t even begin to come close to also trying to keep up on Dr. Casey Fiesler’s lists, Ed Zitron’s newsletter, and the running “To Read -AI” Zotero folder I have going.
4. Have intensive awareness of the tools that have been rolled out on campus, that are being marketed through various university and academic library vendors, and what my students are being told are solutions on Tiktok. Stay ahead of vendors “implementing AI” as black box tools that worsen search results and introduce damage and figure out what needs to be turned off or worked around (again). Needing to spend an extra 8-400 clicks per tool, sometimes every time I use it, to de-activate things is my new least favorite Executive Function Theft example.
5. Understand and engage in advocacy where training sets are stealing from artists and writers, myself include. I know this blog and all of my published papers have been scraped and are being used in clear violation of the Creative Commons NON COMMERCIAL license. I’ve been following the horrors as many internet author friends have learned how many of their book publishers failed to file their copyrights.
6. Work with IRB colleagues as we discuss and navigate the use of these tools and the potential for harm to research participants either during the current project or in downstream harm. The IRB’s primary function is to consider the harm to participants in the collection of data but now we need to have even more conversations about where the data is being stored, who has the algorithm, what happens with the training data, how identifiable the training data is, what the output is…
7. Participate in conversations about how the scraping tools, desperate for more content, are taking down cultural heritage sites and many other small websites on the internet. They’ve knocked out our library digital collections multiple times and I know we are not alone. When the bots show up again and again to pound the servers it forces us to implement added clickthrough’s or barriers, which reduce accessibility and general human usability. For things like our repositories, where researchers might be looking at downloads or views as a (very easily skewed) measure of impacts, we now have to consider the built in bot-factor.
8. Chew on my tongue as colleagues respond repeatedly when I have asked for their personal and well developed and hard earned expertise with “Well, I asked [ChatGPT, Gemini, Co-Pilot]” multiple times in a conversation. I don’t want the prompt output, I want *your* expertise.
9. Sort through dozens of generated near identical resumes and cover letters when trying to hire graduate students. I hire 2-3 new Data Squids a year — a predominantly writing position (which it says in the job ad) — but because the word “data” shows up, students bulk apply to tell me about their indistinguishable programming skills. After HR did first pass, I still had to sort through nearly 100 resumes in August, most of which consisted of walls of text containing programming language names as the students try to battle algorithmic screeners. Although, one student did try to convince me that they had “written data management plans for federal grants” as a primary job duty while a junior analyst at Deloitte.
10. Understand journal policies and funding agency policies around the use of generative AI tools and the associated consequences. Provide education on same. Most journals do not consider ChatGPT to be eligible for an authorship *because it cannot take responsibility for the work.* It remains to be seen whether one of the major LLM providers tries to pursue intellectual property based on prompt outputs.
11. Attempt to pursue grant funding with a critical lens towards AI adoption. Unfortunately my federal agency has been drastically reduced and despite being invited for a full proposal in the spring, study sections did not happen from what we can tell. My collaborator and I are looking for other funding opportunities in between the rest of our work.
12. Navigate discussions with potential co-authors about why I don’t want to use LLMs for any aspect of the research product. These have to happen *early* in the collaboration these days so we don’t waste time.
13. Help colleagues and students confirm that yes, yes that is a garbage output from an LLM citation, no it doesn’t exist, and here are some suggestions for when you now have to talk to the colleague/student/PI about the made up text extrusion. Every academic librarian I know is running into this particular brick wall.
14. Talking to my union about language for our upcoming contract and the FOMO aspects of shoving these tools into our classrooms and required tools like the LMS, resulting in potential advanced surveillance of me and my students and various vendors scraping materials for training without our active consent.
15. Attempting to understand in order to be able to convey security and privacy risks related to both the subscription tools and the open tools. There are a lot of tools being developed. So many. What happens when the tool or the company behind it gets purchased or absorbed or just…handed off with all of that training data? What happens when it includes identifiable human subject data on [insert charged topic here]. I’m grateful my CISO is thinking about this a lot more than I am, but how do I stay abreast?
16. Spending additional time verifying if a text, image, dataset, citation, etc is generated garbage or if it actually exists. I saw a post today about a beautiful little hermit crab on a leaf and how the original poster mentioned being asked “is it real” and what sadness that this is now our first suspicious instinct. And at a time when so many techniques we have been using aren’t available (e.g. what is the source of this material — when everything looks like a chatbot)?
17. Understanding how reliance on the various AI summaries in search engines are introducing potential errors or misinformation. Google trained us well to quickly look at the top result, accept, and move on with our days — but we’ve all seen the ridiculous error ridden summaries or assertions that no, really, that concert wasn’t last year, it’s tonight and that restaurant that closed 6 months ago is open until 9 p.m.
18. Point out the issues with taping and indefinitely retaining transcripts for many of our institutional meetings in order to allow Zoom AI Companion Summaries – especially as these tools are primarily trained on American English language speakers and I have colleagues with many different first languages and related accents. What biases are we further introducing? What conversations does this prevent us from happening?
19. If you can’t be bothered to write it, why should I bother to edit and give you feedback on it? This goes for IRB, colleague manuscripts, peer review. You’re asking for me to spend my time and effort cleaning up something and making it presentable rather than you doing it to start with.
20. Worrying about helping both me and beloved family members navigate “ambient listening” in medical providers offices when I recognize that various clinicians may start firing patients who reject it due to the sensitivity of health discussions and errors we know come along.

I was wondering to myself the other day what else I might have accomplished this year if the above list wasn’t taking up so much time and brain space. What would it mean if I could have given a lecture instead best data practices or gotten to read more more niche history articles? How could we have actually moved machine learning or NLP forward if we weren’t battling the continuous hype? What trust have we lost in potential useful tools? Whose voice or discovery have we lost because of racing goals towards “more productivity” and “faster outputs” which prioritizes volume at all costs?

When we study history — at least as a child studying history– I often wondered how it was that we lost knowledge or a desire to learn. How did “The Dark Ages” (a term now less popular but still a useful shorthand) occur following the fall of the Roman Empire? Why did people abandon literacy?

As I watch people sprint to abdicate their critical thinking skills to “but ChatGPT said!” I wonder less. They offload curiosity, investigation, even their own unique voices to a dressed up autocomplete which is untethered from abilities or responsibilities like telling the truth or fact-checking. As Dr. Ben Zhao pointed out in his keynote to Open Repositories last week, the hallucinations aren’t a bug, they’re actually part what you want the tool to do.

In some ways it feels like a new religion.** Okay, in a lot of ways. You have the various denominations: OpenAI, Anthropic, Claude, Gemini, etc and the associated prophets — predominantly men who have a very pointed group of followers who will hear nothing said against them.

And there are the many acolytes attending to the new faith in the LLM. They are willing to commit their thoughts, emotions, mental health, financial support, job responsibilities, etc all to a surveilling Synthetic Text Extruder, to use Emily Bender and Alex Hanna’s term. Already we see the impacts on the newly faithful:

• People Are Losing Loved Ones to AI-Fueled Spiritual Fantasies
• Concerns grow as dangerous AI trend tears families apart: ‘He would listen to the bot over me’
• Man falls for AI chatbot he created, proposes while partner looks on in disbelief
• Hinge CEO Warns of Future Divorces as Lovebirds Swap Humans for Chatbots in Dating Scene
• Microsoft CEO Satya Nadella Explains How He’s Making Himself Obsolete With AI

It echoes religious seclusion — only they understand me, only they love me, only they speak the truth. One suspects there will be a new monastic order. I worry for those whose mental health will be impacted by increased intrusive thoughts or other harmful behaviors reinforced by a chatbot programmed to affirm (citations deliberately not provided).

But increasingly and troublingly I see LLMs used as a replacement for even a cursory internet search, which might have resulted in a URL or author that could be traceable or provide context as to the source of that outlandish claim repeated from a parody post on a message board. The result instead, presented in a homogenized voice and font, is then presented and accepted without critical engagement – truth enough and likely reinforcing whatever it was the prompter believed they would find.

Thus we get to my opening quote.

Librarianship as a profession often comes with a sense of the religious around it. We have Fobazi Ettarh’s Vocational Awe, where she calls out the messaging we receive that we are to hold our work in wonder and accept poor treatment and pay while pursing a higher calling. There’s even a book called Sacred Stacks, which I have tried to skim and… well, I’ve tried. I’ve got a sixteen bullet point Obsidian note and a correlated Zotero folder started on this from years ago– someone bring me a sociologist, an anthropologist***, and a religious studies scholar to investigate the direct correlations between hierarchies in librarianship and those in the Catholic Church.

With that weight hanging around my work, it feels less surprising then that I feel like a member of an older faith, watching what might be a New Revelatory Faith but might just be a fly-by-night, grift-while-you-can cultish con. (Also if you haven’t read The AI Con by Emily Bender & Alex Hanna, now is a great day to do that.) Who will make money? Who will lose it? Who is being lured in by being “saved” in the new faith where what is saved is very unclear other than perhaps salary lines for the true believers in management? Who will be harmed? What will the faithful be told to believe tomorrow about AGI?

I’m not alone in these thoughts — I’ve seen pieces on Technomysticism that kind of align with it, and there’s this pre-print: GenAI and Religion: Creation, Agency, and Meaning. As usual, I’m working through it in my head and in my essays abroad and always struggling for an elusive conclusion.

*If you’ve been around for a while and read the EFT posts, you know I have two friends abroad who get the most unfettered of my stream of conscious essays. This, by request, is adapted from one of those. (The original is getting dropped at the post office this afternoon and no, I do not maintain original copies.)

**For purposes of this post, I’m leaning primarily on Christianity as a faith because it’s most familiar to me. I recognize that other faiths may have more or less of these features — such as monasteries or the equivalent.

*** Already have my anthropologist actually. She’s standing by for whenever I manage to find time to write this particular piece.

TL/DR: Will the rise of the use of Chatbot style coding to meet basic programming needs and eliminate entry level programmers mean that academic librarians can/will stop teaching the “Intro to [Programming/Python/R/Etc]” workshops that should have been integrated into disciplinary course requirements 10 years ago and/or mandated as computer science dept service courses .

For at least the past 15 years, there has been a steady drumbeat demanding that librarians must teach introductory and occasionally advanced coding skills. Because of where I work, I ‘ve seen this mostly play out on the academic front but having served in LITA and now CORE for a number of years, I’m aware that public libraries have also faced these demands. It is a longstanding gripe of mine that academic libraries have been faced with haphazard demands for coding work with entirely inadequate support or funding to provide regular fundamental programming instruction.

The demands that academic librarians take on this task have paced alongside the fetishized belief that everyone must learn to code and that only code-writing skills are valuable in the workplace and all the good jobs now require coding. The professional community response has been a variety of approaches to fill the educational and skill gap — partially seeking to meet some very real needs and partially because administration was bandwagon jumping. Quite a few academic librarians have chosen or been told to provide intro coding courses not available elsewhere on campus, despite the existence of large and increasing computer science programs with faculty making significantly higher salaries. Justification messaging has included the ideas of “making the library relevant” or “but no one else teaches it” or “but faculty in that discipline don’t know how to code.” This led to the 20teens engagement with the Carpentries programs and other various bootcamp or 2-lecture style events to try and shove especially grad students through foundational skills.*

I’ve seen a lot of this because it’s insidiously crawled into most of the academic data library job ads that I’m still gathering where — in addition to being experts on research data management, library liaisonship, finding data for reuse, navigating data curation, sharing, and privacy, teaching best practices in data visualization, and about 25 other skills sets — data librarians are also expected to teach various R, Python, and other programming skills to students who “won’t get it anywhere else and we have to make sure they are prepared for a data-intensive job.”

That this is still an expectation of academic librarians in 2025 points to a notable issue where faculty and professionals across other disciplines are expecting their students and graduates to have these skills but haven’t addressed it on their end. They have had at least a full decade to train their own faculty, hire faculty to develop and teach integrated courses, and run them through the curriculum committees. And certainly quite a few programs and disciplines have undertaken work to ensure their students are getting this education and training but the consistent rounds of emails I see asking the library where these introductory workshops are being provided on campus suggests there remains a significant gap. I suspect the programs who have not are relying on another longstanding complaint of mine which is that there is a CONSTANT assumption in programming education, including in most intro to programming courses, that you’ve already learned this in some format somewhere else.

Enter ChatGPT, Co-pilot, and the other “synthetic text extruders”** — which, having been trained on Stack overflow and probably all of O’Reilly in the process, does appear to be able to generate “stuff that looks like code.” The current hype headlines tell us it will replace all low level programmers in the next very short while, though I don’t know where we’re supposed to get the senior programmers after that. But the friends I have who do code as at least a semi-regular part of their job responsibilities have described the integration of these tools into their workflows to start the coding process, meet a minimal need etc.

And so the question arose — will the first year grad student take a coding bootcamp next fall? Or will their faculty member expect that they can just chatbot vibe code their way through it? Will the student come in with that same assumption? Will librarians still be getting consistent demands for 1-hour/1-day workshops that get the students entirely up to speed? Will the remnants of our graduate humanities programs drop the coding course because of an expectation that all of the software will provide it for them based on a prompt?

I expect we’re still quite a ways from students actually needing coding instruction and my anticipation is that students will be told to rely on the chatbot and then at some point will get frustrated to the point of being back demanding someone actually teach them the foundational skills again. Or their faculty will get tired of the endless rounds of “looks pretty, doesn’t work” and send them off again for training. At least some of them will. For others it will satisfice and the attendance at this programs will tank.

There’s *plenty* of other work and teaching for libraries to do, but I anticipate at least one library will be faced with “but we don’t need those intro /Carpentries/etc workshops anymore so now the library is even more irrelevant, quick go get 14 NEW skills and be able to teach that.” Standard Tuesday with the Dean just back from CNI.***

An incomplete thought so no pithy closure, but I hope someone does track the progression of this in the next couple years.

*Please note I say skills, not knowledge because these courses are often limited to providing introductory familiarity with an interface and a few simple commands.

** If you haven’t read Emily Bender and Alex Hanna’s AI CON yet, go…go now. It’s wonderful

*** The theoretical Dean, not anyone in specific’s library leadership

As with most technological roll outs, this was a single announcement buried amongst the dozens of emails we receive each day and with a tone of presenting an obvious good. Here was a thing people WANT! and it solved A Problem! The problem apparently being that someone — unclear who — felt like they didn’t have a toy someone else had.

But it’s Approved by the University and Available Now. And that feels very familiar and it worries me. This has been Microsoft and Google’s playbook before. We saw it ~20 years ago and even still when I got to UIC in 2011 with Microsoft Office, back now as Office 365 and then in about 2013? with the rollout of Google for our students. They made it so easy for students and faculty to get access, rapidly normalizing these suites of tools and ensuring that as students graduated they expected their employers would have the same tools and/or that they were willing to pay for personal licenses. We have seen it again in the past decade with the storage offerings from Box / Dropbox / Google / Microsoft. Vendors showed up with these unbelievable prices to offload for us the giant headache of storage and then suddenly the price tripled or quadrupled — causing migration, data loss, and disruption. And students graduating may feel trapped in the ecosystem where their materials already are so they pay for that personal license.

At least with Microsoft Office Suite and the various storage platforms, I saw a specific need being met — though I still have questions about the value of Microsoft Excel’s belief that everything is a date.

In the University Library and especially as a former medical librarian, we have seen this same kind of evolution from “Isn’t It Cool” to “It MUST Be Paid For” with UpToDate — the point of care summary tool that medical students and residents are rapidly inculcated to use, often to the detriment of their ability or willingness to investigate other resources. It’s a pervasive academic medical center integration with a vendor that has shown up with shocking price increases year after year. Last I saw, my institutional cost was 7 figures annually. At one point nearly a decade ago when my then-Dean pointed out that the cost had gone far beyond what the Library budget alone could support, my leadership got emails demanding that the Library faculty all be fired so that access could continue. Notably, you could fire all of the health sciences liaison librarians (9 people) in my building and that wouldn’t equate to the current license cost.

And so I’m worried about the insidiousness of the integration of Co-Pilot. One colleague on Bluesky noted that when they asked “why this tool” at their Co-Pilot rollout — their university IT department’s only response was that they already had a Microsoft contract. Microsoft is happy to leverage these existing and now obligatory relationships for annual Microsoft 365 and One Drive access to add this new added thing whether it is wanted or not and I expect we will see a correlative increase in cost in the forthcoming years as Microsoft attempts to recoup the billions they have poured down the LLM drain. How soon will it be until faculty and students demand that the university provide it and that we just reduce headcount to meet those demands?

Co-Pilot has not been effectively presented with helpful use cases, demonstrated return on investment, actual benefit to students, staff or faculty, or much else other than New! Tool! There have been a few college-specific workshops and links provided to LinkedIn Learning, but notably I didn’t see anything that centered the various forms of harm and homogenization or misinformation these tools have been shown to produce. There’s also a passing mention in our rollout of consulting the IRB in relation to human subject research but there are many other forms of sensitive information that wouldn’t fall under the IRB and should also not be put into these tools (e.g. pre-submission grant application documents). Additionally, faculty, staff, and students are likely to remember “Co-Pilot is approved” and not realize it’s not all versions of Co-Pilot — so we are likely to see sensitive information put into free versions, with the inherent variety of risks downstream.

This is not the first roll out of tools with inadequate safeguards as technology teams race to hop on the current headlines. For us, there had to be a hard pause and review when Zoom AI Companion was rolled out with inadequate consultation about the potential impact on human subject research and informed consent . Personally, I’ve been in conversations with people who should know better about why we cannot allow hospital employees to copy/paste from the electronic medical record into Grammarly. There’s also an entire other essay I have started about how these tools have eaten brain space I didn’t have and have added to my workload without providing any meaningful offset in effort.

But in addition to all the other considerations of ROI, hype chasing, plagiarism issues, intellectual property theft, ecological impact, moderator psychological abuse in the Global South, and inherent English/Western/White/Male centric bias, we should take a step back institutionally and question the long term cost of these tools and how this will likely very soon become another exceptionally expensive annual renewal, especially in these forthcoming days of anticipated austerity.

Academia tends to have limited foci when it comes to mentoring and career directionality — academic librarianship moreso some days. If you don’t want to move across the country every few years, bouncing up a very narrow administrative chain, or you’re not being actively shepherded up into “Great Man” status only awarded to a few, it is not hard to look at the potential decades ahead and worry that you’re headed into stagnation. Further, academic librarianship relies upon but does not particularly value librarians who go deep into their discipline. We award the shiny and shallow, only look at who is on stage giving keynotes versus who is left at home to carry out the work and is not getting those bonus flight points.

Data librarianship remains still relatively new — though colleagues will be the first to remind me that Bill Michener was publishing about this work in the 90s — most of us who are the seniors in the field are only just over a dozen years deep into data management plans, sharing datasets, attempting to explain the finer details of curation, and the slog that is attending to governance. 12 years in the academy is a blip and yet where data was the New Hotness even five years ago, it now often feels like the Old and Busted (Trusted).

To this end and with multiple mentees asking me as they made their own significant transitions “well, now what do I do?” I realized the question was there for me as well.

I’m not job hunting. The invitations to apply that have come my way have predominantly been from institutions that have done very little to build or support data services and the idea of starting entirely over, moving across the country, being given at best less budget and resources than I have now, taking a $30-50k pay cut (one of the more memorable suggested salaries), give up tenure, and having to carve out a niche in a place where my physical autonomy was being drastically restricted just has not appealed.

And I’m not interested in traditional academic library management at present either. With a half dozen graduate assistants, mentees, direct reports, and research undergraduates needing my time, the idea of moving into an associate dean role where I’d spend even more time managing personnel issues isn’t especially appealing. I’d have to step away from what little direct data work I get to do now.

Add that I am on my tenth or eleventh different set of reporting lines in the past decade and a half. I have almost never given my annual review to the same person or combination of individuals for two years in a row, which has added to the disruption and meant that for the majority of my academic career there hasn’t been any kind of consistent focus from a manager. I come up with my own goals, make sure they happen while navigating whatever else showed up this year, report them out to whoever is looking at this year’s final list, attempt to not add too many more goals for the following year, and move on with my day and ever expanding To Do list.

Similarly to most institutions, the university I’m at provides only limited opportunity for professional development and the majority of it is targeted at individuals climbing the administrative lines within their colleges. Our professional organizations similarly struggle with any kind of guidance beyond initial career support or admonitions to just do more service to the organization and won’t you take on more mentorship?

I like mentoring but not when it only works in one direction.

But what comes next and how do I prioritize myself and challenge myself to look internally?

For me the answer was coaching.

There are coaches who are around the academy. The NFCDD sends the proscriptive and oblivious emails weekly, blithely assuming that all faculty work only 9 months, have the summers and spring break off, and are allowed to control their schedules in a remarkably unrealistic way. Oh and that we all write in a very prescribed way that has never aligned with how my head works. Add to that the pricing for their workshops and I knew that was off the plate.

There are many coaches focused on helping faculty to transition to industry. So many friends and colleagues have utterly burnt out over the past five years and are tired of the endless calls for efficiencies, the non-existent raises, the lack of grace from administration and from students where the demands for endless extra giving of time and ourselves have yet to decline even close to the unsustainable heights already demanded in 2019. But that also wasn’t what I was looking for.

There are also a few coaches I know who do work with academic librarians. And at least two of them I knew I would find it potentially conflicting because of their closeness to people in the field. I didn’t want someone whose focus on me would be tempered with “oh but He is Actually Really Great, if only you saw it” when I’m griping about some over glorified bro breathlessly spouting useless nonsense about AI in Libraries.

I wanted someone to focus on My Job Stuff with Me. Who would help me set the time aside and talk through big goal questions and lay out some strategies and ask questions. I needed someone who would respect that I like my job and the majority of the people I work with and that I believe in my institution’s mission. I wanted someone who would respect that my work is a priority for me and will continue to be.

And so I turned to Sara Wachter-Boettcher’s team at Active Voice. I really enjoyed Sara’s book, Technically Wrong, a few years ago and her podcast No You Go / Strong Feelings has been a comfort listen of hearing your ambitious and cool friends who work in areas other than you chat and discuss big questions. From her newsletter, I knew Sara’s company offered coaching. And I also knew that Sara and I, while we could probably have an incredible time over a cocktail and absolutely shredding patriarchal stupidity du jour — were probably not a good match for the kind of coaching I was looking for.

But I knew she had other coaches on her team and so I reached out to them with a very honest email of what I was hoping to find and that’s how I got paired with Jen Dionisio.

Jen and I started working together in late 2023 and we met every two weeks for about four months, moving then to monthly. We took a break over the summer and now we’re back as I reset for the upcoming year.

We’ve talked at length about ambition in the workplace and the current trend towards distaste for that. We talked about how most of the guidance for women my age is focused on balancing childcare with the rest of your obligations and how that erases or diminishes the challenges that may be faced that aren’t related to parenting. (The “Oh you don’t have kids so YOU can do this extra service” penalty) We have navigated together how to separate things that are Mine to Carry and when sometimes it is something that others need to address for themselves at work and I cannot be the one to fix it all for them. She gives me homework — some of which has been to read through the various self-help/ job related / no really you aren’t bonkers it’s society right now books that have been piling up in my office.

And we’ve talked about what I want to do. What does 3 years from now look like? For the precious hours — Jen and I focus on my professional self and what comes next for me.

I told her a couple of months into coaching that it’s like having a really good manager. Someone who is actively invested in my success and in helping me get barriers out of my way. Jen’s not local and she can’t actually change things at my workplace. And she doesn’t know my coworkers, so instead she gets the glimpses only from my view and we talk about the stereotypes of office and professional politics that transplant amongst nearly all workplaces.

And we keep looping back to my bigger goals, which have been absent from many of the discussions with library managers I’ve bounced between. If the goal is the Annual Review and Tenure — that’s often the only things discussed. What does success look like at a more abstract range? What would I actually like to see happen and can I start putting those stepping stones in place now rather than getting totally knocked off course — pending things like the Nelson Memo showing up and changing federal policies on me again. Working with Jen has also given me frameworks to push against and to recognize what is General Work Stupidity versus what may actually be personal and where I need to take action or advocacy.

I’ve been writing and speaking over the past 18 months about the issues related to our institutions and organizations and the absence of mid-career support and that’s a research topic that I hope to dive into further in the coming years. But for those who are waiting for me to get your institutions on board, a bridge step might be coaching. Having someone who knows *you* and enthusiastically wants to help you to succeed might just be the mid-career support you need. A+ Do Recommend.

Please understand that I have comprehensively raised these concerns in writing and synchronously with MLA staff. This is not a surprise to them and MLA has and will likely continue to ignore me. Though from what I can tell and will describe below, I appear to not be the only one who has brought these issues forward. MLA seems uninterested in making modifications to their contracts that would better serve the members or other solicited contributors whose intellectual property and knowledge comprises their extremely expensive annual conference* and other continuing education and professional development.

So what am I talking about? This following are two screenshots from the call for proposals for the MLA 2025 conference. The alt text has the full text copied. I have a full page capture in Zotero at the time when I start this draft in the event that they do modify and improve it.

“You hereby grant to MLA the following rights to any written or visual material submitted in connection with your presentation(s) at the Annual Conference in any and all media or form of communication whether now or existing or hereafter developed:”

Any and all forms? Do you suppose this includes feeding posters, slides, or recordings into an LLM? Oh and they can repackage it and resell it without telling you indefinitely? And they can use your image to promote the organization — whose images will be used and where? And any other publication they want to sell? Or they could repackage your presentation and sell it to a secondary non-librarian audience, again, no honorarium for you, possibly not even an additional CV credit. All this for you to then pay to attend and present at the conference.

This year, in 2024, I noticed this addition of the “based on past practices” to the Call for Conference Proposals. I can’t say for certain when it was added, if someone remembers it from 2023, let me know. It’s a meaningless addition because the contract continues to be overly broad and demand rights that they do not need to conduct the conference or other continuing education. MLA’s speaker honorarium, when offered, certainly does not reach what anything close to market rates for work-for-hire, which is how I could read this contract.

MLA is overreaching in this contract and it needs to be revised. This especially causes tension for medical and hospital librarians who are obligated to participate in order to maintain AHIP** status or pursue promotion, tenure, or new jobs. I am reasonably confident that many librarians who want to share their research, their case studies, and their interesting things are not anticipating that the organization claims permanent access to their *likeness* and the ability to indefinitely resell their work.

The organization could update the contract to get what they actually need from presenters in order to advertise the conference and have a limited option for recording/sharing afterwards. No need for any existing or hereafter clauses.

N. Sims, copyright librarian and lawyer extraordinaire, has written on their blog about the need to consider against whom your lawyer is being protective and combative and this is *also* something I have shared with MLA over the years. The MLA lawyers appear to be so determined to get things and control them for MLA that they have crafted a contract that isn’t in the favor of the membership. I believe it introduces harm because it normalizes this overreach.

While I have been an MLA member in the past, I stopped a number of years ago, though I continue to be a member of MLA Midwest Chapter. The organization wasn’t my professional home and the costs were outsized compared to the value. Notably, MLA staff have suggested several times over the years that despite not being a member I somehow still owe the organization free or deeply discounted contributions of my time and expertise. The journal, where I did serve on the editorial board, is still a place I will consider for article submission and that I will review for but that’s the extent of my engagement at present.

I hope that MLA members will call for better contractual language from the organization that is supposed to be serving them. They deserve it.

*2024 registration for a non-speaking member was just under $800
** Academy of Health Information Professionals is an add on that requires continuing education, doing service for MLA regularly, etc. It’s a requirement for some medical library roles — usually in administration.

Also, a starter caveat. Below, I’m about to be highly critical of surveys that were distributed to librarians. This is my discipline and, having written my own not very excellent survey instruments, I feel qualified to have loud opinions about it. Librarianship, however, is not alone. I’ve seen the instruments that come to my IRB and from students and faculty sent to and from a variety of disciplines. Poor survey design and re-identifiability is an issue we all need to address. Fix your own work.

If you have followed me on Mastodon in the past two years, you may have seen a post come across where there’s a weird list of Today’s Survey Demographics. I started this one day when yet another survey hit one of the myriad professional listservs I lurk on, in order to be publicly annoyed at how quickly I could be re-identified from the “entirely anonymous” survey that had been distributed. What else is social media for? Then about 15 months ago, I started capturing the survey instruments to see what the demographic question trends really were and the variety of risks surveyors were introducing.

At point of writing I have summarized information from 23 surveys that were fielded between July 2023-November 2024. These surveys are primarily aimed at academic librarians. The majority of them sent recruitment somewhere on ALA Connect. This is, indeed, a convenience sample of surveys –but considering the surveys themselves were seeking a convenience sample, that seemed appropriate.

I have coded the surveys for the following:

• Survey Topic
• Does the instrument include the informed consent
• Anonymous / de-identified in informed consent
• Types and number of demographic questions
• Clearly marked required questions
• Appropriate skip logic used
• Demographic or content questions first
• Did I notice a significant number of leading or otherwise biased questions?

Survey Topics:

As you can see below, topics varied all over the place, but gives you an idea of what is happening in librarianship. And a few of these survey authors suddenly know who wrote all those “hey, this is someone looking at your survey” or “please delete this row, it’s someone looking at the questions” every time they came across an “other” box. For the record, I did not meaningfully participate in any of these surveys.

Informed Consent

Only two surveys didn’t have their informed consent with the survey instrument itself; one had it in an email only and the other I’m not sure what was going on. Over half of the surveys asserted in that informed consent that participants would be anonymous. Only one survey explicitly noted that they were going to ask for identifiable information, which included the name of the Library.

Type and Number of Demographic Questions

My specific interest was the variety of survey demographic questions being asked. Before I started coding, I made a list of about 20 areas based on what I could remember from my Mastodon posts. The was edited to the final list of 25 categories, listed below here by frequency.

The number of demographic questions spanned in the identified categories between 0-14. The most common was between 4-6 demographic questions, though there’s very little consistency beyond that. Shout out to the one survey that had 0 demographic questions at all.

Despite my attempts to come up with an exhaustive category list, I missed a few. Other demographic questions that showed up included: ARL Status; repeated and expanded questions about years worked/where/num of librarians; PT/FT; Name of Institution; Department Names in Your Library; Work in a Specific State; MLS specifically y/n; Parent’s Highest Degree; Household Income; Childhood Household Income; Country; Exact Job Title; Library as First Career; and STEM Degree.

In most instances it was entirely unclear how the authors chose which demographic questions to ask, in what order, and what they expected to learn from their respondents. But the survey that asked the most had a total of 18 demographic questions — 14 from the categories above and then an additional 4. I cannot imagine the response rate for the survey was going to provide nearly enough responses to get any kind of useful breakdown.

When Did They Ask the Demographic Questions

When we did DataDoubles, our survey had an extensive set of demographic questions. We asked about race, ethnicity, political affiliation, etc. because we were indeed hoping to understand the full variance of students across our 8 institutions. And all of those questions were at the very end of the survey. I’d have to go back and confer with the team about the percentage of students who dropped off after the demographics — there certainly was a significant amount. But we still successfully got their thoughts on privacy and learning analytics first, which was what mattered the most. Not so for most of the surveys here, where 73% of the surveys started with the demographics before moving into the subject content.

Other Challenges: Required Questions; Skip Logic; Biased or Leading Questions

As I read through the surveys it was frequently unclear which questions were required. Some had all questions required, other partial, most none. But in 87% of the surveys is was difficult to tell what was actually required of the participants. An additional challenge was a not infrequent lack of skip logic. I would be asked a question where the answer was no and then scroll down to four more questions that assumed I had answered yes.

Finally, about a quarter of the surveys were loaded with biased or leading questions. In one, they asked about “benefits” that were internal and focused on an individual while “challenges” were focused externally on systems/groups. There was a lot of participant cueing related to buzzwordy concepts and it was very clear that the authors were aiming to assert that the Vast Majority of Respondents Want Shiny. (This was particularly so with the AI surveys.)

So Why Does this Matter?

Privacy is dead, right? Anonymity can’t happen anymore, we’re all too online and re-identifiable? No one will actually try to re-identify you so why are you worrying?

No, privacy isn’t dead and it’s a lie to assert anonymity in these surveys. In a number of these surveys it would be trivial to re-identify me. If you ask for the US geographic location I work in, if I’m tenured, if I’m at a public institution, oh and what specialty of library I’m in — well… there’s not all that many tenured data librarians. And to use people with power: my current dean is a tenured Black woman in the Midwest and the current ACRL President is a tenured Asian man who is Dean out in the Southwest. Take the gender/race/geographic location/ and tenure status and tada … you’re looking at a tiny pool of colleagues. This has significant potential to re-identify colleagues who aren’t the majority demographic pool and potentially introduce harm. And some of these surveys are very clearly touching on sensitive topics like mental health.

I also worry as a data librarian because our journals, funders, and disciplines are expecting more data sharing. It’s not only you that I have to worry about, it’s everyone else who downloads the dataset from the data repository and decides to see what they can reverse engineer or whether they can sort out that one anomaly. We’ve seen our own bad history of data security handling, Kristin Briney and Becky Yoose have written about it and there was the dataset last year from Code4Lib Journal that had *all* the patron information. And while most of these went through some form of IRB review, it’s clear from just this sample that IRB isn’t sufficient to understand the re-identifiability risks.

And I don’t see the point of most of these demographic questions. It generally looks hasty and rushed — oh let’s slap some demographics on there to look more research-y. We’ll be able to do categories on it in our findings! Nevermind that we don’t have a good idea of the total pool or how representative it is. Nor that it’s probably not all that different to talk about General Topics at a large M1 institution versus an R1 without much more significant investigation or comparison of services.

So You Have Best Practices for Us, Right?

If you’d like best practices for library survey design, then I encourage you to investigate the literature and discover who is writing about this. Pay an expert to review the questions before it goes out — your campus may have that resource available to you. But I am not that expert nor do I aspire to be. I will leave you with my typical suggestion to read Dr. LaTanya Sweeney’s work on k-anonymity and the following suggestions as a starting point:

I look forward to seeing your improved surveys.
(P.S. If anyone reading this *is* an expert in this area and would like to be added to this post as someone who does consulting in this area, please send along your info.)

I wasn’t entirely sure what to expect when I put it out in the world. I was writing as many of us were leaving Twitter, which meant that the mechanism where I was used to getting the most uptake for my blog posts was gone. Mastodon and Bluesky weren’t and still aren’t quite where I saw the same kind of Escape Launch Virality, which is overall more of a good thing than a bad thing. But the post started making the rounds and getting good response.

And then a couple of days later we were scheduled to fly to Florida and on the way I got a text message from Griffey: “Did you know you’re on Hacker News?” I did not, but I spent the next several days trying to balance doing things with the Philosopher’s family and frantically checking my stats and comments.

Executive Function Theft, as I noted in my entry last year, becomes so obvious when named. I have heard from so many people whose entire response is “Oh, oh is THAT what it’s called. Yes, yes that.” It has given a phrase for a feeling, an issue, a problem that we can see but couldn’t succinctly describe. I’ve heard from colleagues and friends whose managers were committing EFT, endless healthcare examples , and dudes (mostly cishet) assigning 10 step tasks to their partners and demanding to be lauded for doing 1/2 of the final step.

Criticism seemed to center around themes of:

• You should just automate all those things. — My guess is if I asked how to automate many of the tasks I described, it would come back to me with “pay someone pennies to do it for you” which has all sorts of ethical issues or loses the actual care/community parts of care tasks.
• “I don’t care about those things, so you shouldn’t either. ” — Odd, yes, I do care about getting reimbursed for my medical expenses.
• But CAL NEWPORT!! — Cool, that must have been in the part of the book I couldn’t make it through.
• “You’re clout seeking” — By writing an essay on a not especially high traffic blog. Right, yes.
• Oh and there was one guy who tried to tell me I meant the opposite of what I wrote. There’s always one.

But most readers have been great and I’m grateful that have had such a positive reception to this idea and concept. Seeing EFT move into the wild and beyond my own networks and the immediate boost when it hit last August has been fascinating. I hope it will continue to grow and to be useful to new people.

And yesterday, I was delighted to see that maybe even the federal government is getting in on reducing Executive Function Theft! Make sure you give them additional suggestions where paperwork reduction and process simplification could help all of us.

May we all have reduced EFT in the future!