Hi Eric, Steve, Roy,

Chiming in, I wrote an immediate response to the Parable AI 'hype' story, which is creative writing.

But I want to emphasize and echo a few pieces of Steve's very good response on this piece. Thank you Steve.
The three buckets are very important and accurate. I appreciate that librarians are more than ready to grapple with the fact that technology comes at a cost, and we're in an era when environmental costs are very real, and it's ok to measure the use of a tech gadget against that cost.
I appreciate Steve's sidenote that your work with an LLM was not 'quick and easy' — sooooooooooooo many people still have trouble formatting a microsoft excel sheet for export to csv. Digital literacy divide is very real. The danger of a tool like "AI" being a black box to those folks is a real risk to the research landscape, for reproducibility, transparency, and hasty conclusions. To say nothing of the general information landscape, where librarians are very invested.

And just to add to that, the fact that you assert what the LLM output gave you is true: you have a k-18 educational career without AI behind you, I assume, perhaps you had AI for undergrad or grad. So you can in fact reason with whether the output was correct. If you had had to fact check it, how would you have? If you had been asked for a data sharing agreement from every participant on the CODE4LIB listserv to satisfy an IRB requirement, could you have? If you wanted to attribute the direct quote, "Annoying data format" could you have? These are the steps of the research process that AI obscures, without rigorous critical engagement at least. So, librarians are standing for rigorous critical engagement, which sometimes means not using a tool that cannot be verified.

All this to say, if someone's instinct is at all to hesitate in the face of AI, I urge us to listen to it.

With respect,
Adrienne

RE: A parable, I pose another point of view

You are seeking a book, or something. It's important, so you think it's in likely to be in a book.

It is a truth universally acknowledged that where there are books there are librarians, so you seek the library.
And there you are directed to the right reading level, and the right aisle, and the right words.

At least statistically. A 28% chance of error<https://urldefense.com/v3/__https:/www.omnicalculator.com/reports/orca-ai-benchmark-2026-update__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3vjGqlaX$> means naught in this age.
Technology tends to change quickly, depending on who can make the most money rapidly off it and feed the next hubris. Humans change more slowly, which is perhaps why research shows<https://urldefense.com/v3/__https:/news.stanford.edu/stories/2023/04/technology-might-be-making-education-worse__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3mnaeVkN$> education systems work best with less technology.


But instead, what you're seeking is not there. Because the genres have been chewed up and spit out, a stochastic repitition of what has come before. The churning of the algorithm simply reproduced the past: bias, harm, and knowledge in all. But this is a new thing.
Climate fiction is not science fiction, it is not science fantasy, and it is not science.
Romantasy is not romance, or fantasy, or is it?
Project 2025 is political systems. Or recent history. Or wait, that search term doesn't even come up in this system?

The person running the machine that spits labels did not know all of these foibles, how could they, they've been asked to know everything about the whole building for years.<https://urldefense.com/v3/__https:/crl.acrl.org/index.php/crl/article/view/26106/34029__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3vsNQIDi$> <https://urldefense.com/v3/__https:/crl.acrl.org/index.php/crl/article/view/26106/34029__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3vsNQIDi$> Their expectations are sky high, while their time is the same as ever: a 12 hour day, since the union went bust at behest of the same conglomerate that sold us the AI labeler.

"But you have AI to help you now, right? Efficiency is progress, efficiency is required, efficiency is best."
And so context is lost<https://urldefense.com/v3/__https:/openenvironmentaldataproject.substack.com/p/weird-sight-against-epistemic-dominance__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3s19_Cz3$>. Choice is an illusion. Solidarity with the data workers piad even less than you<https://urldefense.com/v3/__https:/dl.acm.org/doi/10.1145/3772318.3791097__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3jICMeKW$> (painful, I know) is 'not possible' because AI is 'inevitable' so, so much is lost.<https://urldefense.com/v3/__https:/substack.com/home/post/p-193721103__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3tPEhw-B$>

Meanwhile the prediction that AI would make things easier is slowly being revealed as false. The proven record of harm of AI reducing job satisfication<https://urldefense.com/v3/__https:/fortune.com/2026/02/10/ai-future-of-work-white-collar-employees-technology-productivity-burnout-research-uc-berkeley/__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3oys6LQ1$>, creating more work<https://urldefense.com/v3/__https:/futurism.com/artificial-intelligence/ai-forcing-employees-work-harder__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3p2ab5cP$>, critical thinking skills<https://urldefense.com/v3/__https:/www.mdpi.com/2075-4698/15/1/6__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3qWU17lO$>, and making people feel detached from the meaningfulness of their work<https://urldefense.com/v3/__https:/evilmartians.com/chronicles/ai-assisted-engineers-are-burning-out-is-this-fine__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3vxVCbC5$>. The idea that AI freed us up to do more work with humans was unable to come to fruition, dying on the vine of budget cuts, burst dreams<https://urldefense.com/v3/__https:/archive.is/sbuaE__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3qOvU5G_$> and campaigns that bothered not to hide their intent<https://urldefense.com/v3/__https:/theconversation.com/how-transparent-policies-can-protect-florida-school-libraries-amid-efforts-to-ban-books-269769__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3oPcE-9U$>. Instead AI found the demons<https://urldefense.com/v3/__https:/bookriot.com/the-greenville-eight/__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3mo4kHF7$> in our closet <https://urldefense.com/v3/__https:/www.inthelibrarywiththeleadpipe.org/2026/seeking-approval/__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3gHe9PyZ$> and let them out, let them loose<https://urldefense.com/v3/__https:/www.inthelibrarywiththeleadpipe.org/2026/librarianship-and-maori-knowledge/__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3jSMHHbG$>, let them play in plain sight<https://urldefense.com/v3/__https:/www2.archivists.org/groups/human-rights-archives-section/librarianship-at-the-crossroads-of-ice-surveillance-by-sarah-la__;!!CGUSO5OYRnA7CQ!Y2isGFypGA5sSQxTIHgJXV7tpYrd3M1yLO3H0vkbLBwBiEwbcfmXTWjL2YfqjboGSiqOaXa1AdwGNmoDOSLkRZ-YAVnz3ldNtVkS$>,
for want of a little  more efficiency.
For want of a leisurely future.
Because the sower threw seed hastily.
Instead of taking care of the soil.


Adrienne Canino (she/her)
Data Librarian (MLIS)
TetraTech Ocean Sciences
Anchorage, Alaska
Book a 1 on 1 meeting with me for data management Q&A, or Show and Tell<[log in to unmask]" target="_blank">https:[log in to unmask]>

Data Management Practices<https://www.axiomdatascience.com/best-practices/DataManagementBestPractices.html>
Research Workspace Tutorials<https://researchworkspace.com/help/NavigatingTheWorkspaceTutorial.html>