I am very much on board with Tara, as it is now I would not vote for it. Until there are generative AI models trained on entirely ethically obtained material (and material that publishers provide without author recourse is NOT ethically provided in my opinion, even if it may be legal), and until there are AI models that are not predicated on extreme environmental damage and supported by hoards of underpaid workers in 3rd world countries in Africa, I cannot in good conscience advocate for the use of AI in libraries. There are some institutional AIs being built which seem to have addressed a lot of these questions, and also the question of AI hallucinations (I am coming from an academic background where I too have had patrons asking me to find citations which never existed), but they are few and far between and mainly only useful at this point at the institution which created and trained them on internal documents.
I do agree that there DOES need to be concrete training on navigating the AI being forced onto us by large corporations, training that ties into information literacy and how to navigate the often dangerous information that may be being spit out at patrons. I also think there should be more training on the things that pre-ChatGPT AI were already doing, and how generative AI works, and the difference between AI predicting likely sentence structures and idea-clusters, and actually genuinely creating something new. I think a lot of people believe AI is genuinely creating new knowledge in a way that it simply does not. I also would like to see more explicit instruction on the kinds of bias and bigotry that AI perpetuates due to its training models on flawed data. There also needs to be training on the ways in which over-use of AI can actually harm your own critical thinking skills and prevent you from learning thinking and reasoning skills for yourself.
I would like to see ALA leading the way on advocating for more ethical AI models to be made available. As it stands now, the Vision Statement is a contradiction, because there simply is no way to ethically use the current generative AI models available to the public. AI may become extremely useful in some situations in the future, particularly in medical fields where it already seems to be having some success, but I would not like to see ALA falling into the trap of encouraging whole-sale adoption at this point in time when that potential has not been realized yet. I think the current trend of uncritical adoption by educational institutions is dangerous and counterproductive, in particular. Let's not do that.
------------------------------
Kestrel Ward MSI/MA
they/them
Digital Services and OER Specialist
Florida Virtual Campus Library Services
https://libraries.flvc.org/------------------------------
Original Message:
Sent: Mar 05, 2025 11:57 AM
From: Tara Brady
Subject: ITAC: ALA and AI Vision Statement Presentation
I guess my first-glance issue with all of this is that I'm not totally sure how anyone can implement it in practice. It seems to be strongly advocating for the use of AI in libraries, but then tells us for example to ensure "that AI is not used as a replacement for accessing information from diverse viewpoints, nor to infringe upon intellectual property of others. " The models are what they are, right? They're essentially designed to reinforce the systemic biases of the data they're trained on. How do we avoid that other than by just not using them? Many of them - the most easily accessible ones in particular - are essentially plagiarism machines, regurgitating content they've soaked up without the authors' permission and making the still-untested claim that it's all fair use. How do we ensure we're not infringing on intellectual property other than by not using them? Similarly, we can 'advocate' for them to be 'energy-efficient' all day long, but currently (with the notable exception of one newer model that I'm not sure has even been fully rolled out and tested yet) they're very much not. They're infamously wasteful.
I think that we are absolutely in need of guidelines for best practices on using AI in libraries. We do need real meaningful guidance about when and why and how (and yes, even if) to use AI, and about how to give our patrons realistic and useful instruction when they come to us with questions about it. But I don't know that adopting this vision statement is the right start, because it does come across as saying that we can and will use it in ways that I'm not sure we currently can, at least not without violating our core values.
Sorry to come across as pretty negative on this but I'm coming at it from the perspective of a public librarian who's had patrons come to me looking for citations and even fields of research that don't exist, because the chatbot told them they did, and who has had patrons ask me to 'proofread' LLM-generated essays that as it turned out they hadn't even read themselves, let alone written! Let alone all the dangerously bad answers it often generates. So I'm a little bit jaded about this stuff. For me, what I need from ALA is real, practical guidance on how to educate patrons about AI responsibly, and how we are going to influence the technology's development in a way that is less harmful to users and to the planet.
Thank you, genuinely, for getting a conversation started. I don't think I'd vote for this, personally, but I hope that in time it can get us moving on something that will be very valuable for ALA and its members.
------------------------------
Tara Brady
Queens Public Library
She/Her/Hers
Original Message:
Sent: Mar 05, 2025 11:03 AM
From: Charles McAuley
Subject: ITAC: ALA and AI Vision Statement Presentation
Greetings Council! The Information Technology Advisory Committee has drafted a "Statement on the use of Artificial Intelligence and ALA Core Values" as a vision statement from ALA to all member libraries for their use of artificial intelligence, as they begin integrating it with their systems.
We would, very much, like to ask for committee review and endorsement. Following revision based on review, it is our hope to bring the resolution to Council at the ALA Annual in Philadelphia this June.
Thank you for taking the time to read over the attachments, which include the Values Statement as well as a Resolution for potential adoption.
All feedback is welcome. All comments are appreciated.
Please feel free to e-mail me at any time.
All the best,
Chip McAuley, PhD
Chair, Information Technology Advisory Committee
------------------------------
Charles McAuley
Social Media Librarian/Adult Services Librarian
Mercer County Library System
------------------------------