R. David Lankes talks with Santosh C. Hulagabali on the recent developments concerning libraries, librarianship and AI technology.
Lankes is the Virginia and Charles Bowden Professor of Librarianship at the University of Texas at Austin’s School of Information. He is the recipient of ALA’s Reference and User Services Association 2021 Isadore Gilbert Mudge Award for distinguished contribution to reference librarianship. His book, The Atlas of New Librarianship won the 2012 ABC-CLIO/Greenwood Award for the Best Book in Library Literature. Lankes is a passionate advocate for librarians and their essential role in today’s society.
In this interview for Open Interview, Lankes shares his ideas and experiences about dealing with the key technological changes that are taking place in the academic sector. He discusses as how libraries have continued doing good work despite there are challenges revolving around the use and access even in the times of text-generative AI technology. Further, he opines that mission of libraries has to be improving society through facilitating knowledge creation in communities. And this could be possible by MAKE which stands for motivation, access, knowledge and environment.
.
.
• You have extensively written and spoken about new librarianship? For the knowledge of our readers, could you please explain what does new librarianship mean and how to see ‘new librarianship’ in the changing technological landscape?
New librarianship looks at the work of librarians–in a library, or outside of it– grounded in community learning. Librarians are not defined by where they work or what they do: inform, or catalogue, or collect. Rather we facilitate knowledge creation (learning). We may well collect, catalogue and inform, but these are the result of learning. Ultimately it is about the mission of libraries to improve society through facilitating knowledge creation in their communities.
It is also based on how people learn (through conversation) and the nature of knowledge (uniquely human). It is also a strong movement to shift librarians away from reacting to technological challenges to what librarians do to working hand in hand with communities to find opportunity in change, and mediate the unanticipated consequences of technological determinism.
• Everytime libraries witness major technological development– be it the induction of computers in libraries, emergence of library automation software, use of the internet, e-books, robotics, and now AI technology, we write, discuss and debate on existential or survival issues of libraries and librarianship. How libraries are reshaping or sustaining themselves to be relevant?
I would take issue with the idea that every technological-advance leads to every library having the existential debate. While there is plenty of handwringing, there is also a lot of innovation that occurs in these times of disruption. Library automation actually grew out of academic librarians driving the mainframe and database development in the 1970s for example. But I take your point.
First, let’s give folks a bit of a break. I’ve been involved in libraries since the 1990s. In those 30 years, librarians have had to deal with the rapid adoption of the internet, then the web, then digitization, then a mobile telecommunications revolution that pulled the phone (and computers) from the walls and put them in our pockets. In the last ten years, we’ve seen the explosion ofgenerative AI, a global pandemic, and now a weaponizing political ideology. That’s a lot for anyhuman being to deal with.
The key is to stop focusing on what we do, and focus on why we do it. It gets back to your question about new librarianship. Google is only a threat to librarians if you define librarians as search engines. In truth, as Kim Silk points out, Google has been the greatest document delivery innovation ever. The fact that reference librarians can find a source, and send a community member a link to full text is a spectacular advance in reference: all funded and developed by the private sector.
I look across our technological upheaval in libraries and I see a faster and better response by librarians developing. With the rush of generative AI news, I’ve seen a lot more professors worrying about their future than librarians. Instead, librarians are putting their skills to work evaluating these services. Pointing out “hallucinations,” gender bias in responses and so on. Librarians are already incorporating ChatGPT into information literacy lessons. The Toronto Public Library has been offering workshops and classes on AI and machine learning for years now.
I attribute this new proactivity to a shift from functional definitions of librarianship to more fundamental service and community model. Instead of worrying about putting librarians out of a job, we’re worrying about what happens to folks in the community that might be displaced or even abused.
• With AI add-on in many of the tools, how do we need to redefine catering library services and research support services in universities?
We have a long way to go here. One of my hopes is that we focus on the services of our parent institutions. Take an academic library, for instance. We can spend a lot of time doing LibGuides and updating our instructional activities. But the real issue is how generative AI will change the work of faculty. We have questions like: Will students be handing in AI generated assignments? Is that cheating? How can one better craft assignments to incorporate these functions and focus on improving students’ critical thinking?
I was in a research team meeting the other day and we were talking about a study. We kept roaming around methods and data, but the key question was “what is the research question?” Aquick trip to ChatGPT led to a really solid draft of a research statement to advance our conversation. Likewise, I was putting in a fast turn around proposal on staff training. We needed learning outcomes, and a visit to ChatGPT handled it. Sure, it needed human prompts, editing, and verification, but it made it very clear that my life as a faculty member might be changing. I could use guidance from the library.
What happens in our towns and cities when AI is generating new polices and legislation? Librarianship is the best equipped profession to use and shape generative AI. We’ve thought a lot about sourcing, provenance, intellectual property, scholarly communications, spam, etc. The communities we serve, on the other hand, have not. They need us now.
One of the key services we should be thinking about is credibility. How do we work with our faculties and students to ensure credible information is used in research and assignments, and how do we ensure the credibility of the faculty beyond the institution? How do we establish trust and credibility in sources that do not rely on human authorship? Librarians think like that. It is a very different word from scholars initiated in invisible colleges and their disciplines.
• It has reference with one of your works, i.e. “Eulogy for the Information Age: The future is impact not access“. How do you see both impact and access in the context of AI technology in libraries? What has changed and what hasn’t?
Let me do a little terminology parsing here. The term “AI” is very general, but until about six months ago, it referred mostly behind the scenes use of machine learning to optimize data. Using data for recommending materials, showing posts, charting fitness goals, suggesting music, etc. This view of a sort of invisible hand or “algorithm,” has some potential, but limited potential, in many library applications. Why limited? Because libraries do not do rich data. Take things like bibliographic metadata that for centuries was taught to be spares, but precise-just a few terms, but controlled terminology. We didn’t want noise in catalogue records like transient use data or folksonomies. This is why Google and search engines just couldn’t really use this data effectively. This is why a book in a library collection never showed up at the top of results. It’s also why discovery layers are good, but limited.
The other reason is that due to libraries fighting to preserve the privacy of our communities. We simply don’t collect all the data we could on materials or those using them. Our ethics are constraining wholesale transformation of patrons into data points. So, recommender systems (you liked this resource, you might like this one) will always be limited.
This deliberately limited data landscape makes the use of the invisible hand of AI limited. The way that libraries have gotten around these limitations is in precision, but also in moving to community engagement where there are rich human to human interactions. Where the librarian doesn’t sitand collect data from people, but forms a relationship and broader knowledge of the community served. This shift of focus from transactions and defining library success by document delivery to relationships is what I was talking about in the eulogy.
Now, however, generative AI is shifting this world in one very important way. As libraries have moved to more relationships, they have also shifted from delivering existing materials to creating new expressions of knowledge. So, will libraries have to figure out what we do about acquiring stuff created by AI (books, articles, etc.)? Yes, but what about the ability for the average person to use generative AI for their own creations– write a book and need illustrations? Librarians now have new tools to help folks make their ow content. Need a theme song for your new podcast? Need a voice to narrate your virtual research poster? Generative AI might unleash expression, not replace it.
Now, there are huge issues here. From AI image creation tools being trained on copyrighted images, to a tendency to generate stereotypical imagery. But those issues will not be addressed by AI, it is a facilitated process– facilitators like librarians.
• Copyright, fair use, and open licenses are well concerned with the libraries. Now due to AI generative literature, what challenges and issues the libraries need to consider?
In the US, the copyright office in the Library of Congress has determined human agency is necessary for copyright. In essence, only people own their creative expressions. This, to me, is a big opportunity. We’ve already taken massive steps forward in the open access world by making publicly-funded research freely available. If we now say that AI generated abstracts, and summations also can’t be locked down, then we have a powerful means to distribute knowledge beyond the university, and beyond the global west.
Our goal in academia has always supposedly been the wide distribution of findings. AI can help us do that quicker, using more media (images, videos, narrations). That’s good. Do we need to make clear that stuff has been reviewed and documented (what is created versus what is actual imagery or data)? Absolutely.
I am very excited about what Khan Academy is doing. They are using generative and conversational AI as a tutor. Learners aren’t simply given an answer or list of relevant citations. Now generative AI can be trained on excellent pedagogy, and help lead the learner who can’t afford one-on-one help. Not to replace great teachers, or professors, but to augment them in a way that is economically unfeasible now.
There are real problems with how large language models are developed (environmental impact is huge), but it is not too far from existing legal precedent on search engines. Except in the visual and auditory world. Here there has always been a debate between influence and theft. The old joke is- – if I copy your work, I’m influenced by you. But if you copy mine, that’s theft.
This is going to force a real reckoning in copyright law, as well as art and design. In many ways it throws back to the development of Hip Hop and the sampling revolution in music 30 years ago. We have a long way to go.
• Privacy, biasness, secrecy and transparency- what do they mean to Librarianship when it comes to dealing with AI?
Exactly what they have always meant. A librarian is defined by three inter-locking components: Mission, means, and values. I’ve mentioned the mission before– improving society throughfacilitating knowledge creation in communities. But that’s not enough. Teachers, publishers, and more could share that mission. How do we fulfill this mission? It’s by MAKE that stands for motivation, access, knowledge and environment.
We understand and build on a person’s ‘motivation’. We provide ‘access’ not only to stuff, but other people with a shared interest. We provide basic ‘knowledge’ to facilitate these interactions (language training, background material). And finally, we build ‘environments’ where people can feel safe exploring dangerous ideas. This might be a building, or a forum, or an email chain.
Still the means and mission are shared across many professions. So, the last part we need is the values that guide us in how we enact our mission. Google seeks to create knowledge and they certainly provide access, but they collect a whole hell of data to do this in a way that makes money.
Librarians value learning, openness, diversity, intellectual honesty (not unbiased), intellectual freedom and safety.
None of that changes with AI, that’s why I don’t see it as an existential crisis for librarians. That mission is durable and has been for thousands of years, even if the words we use to state it has changed. Those values reach across communities and situations. The specific tools we use to do the work will always change, and are never the sole defining factor of the field.
• How important is it to have a policy-driven or standard-driven approach in place for the judicious use of AI at higher educational institutes? How do libraries need to go about it?
Last spring, I taught the introductory survey course for the information science degree here at Texas. Every Masters student had to take it when they start the program. In that class of 50, there were future librarians and archivists, but also data scientists and UX designers. ChatGPT exploded on to the scene right as the class was starting. Every topic we were talking about from information seeking, to memory organizations, to intellectual property, to credibility was cast in a new light right around us. One week, a student would ask who owns the output of ChatGPT and the next week the copyright office made a declaration that only humans could claim copyright. It was wild. In fact, that last week of class, a vendor that checked student submissions for plagiarism (Turnitin) was looking for faculty to try out their AI detection App.
If you had asked me to write policy at the beginning of the class, it would be radically different than what I would write today. In fact, I asked students to try ChatGPT, and feel free to use AI in their assignments. I was curious how they would use it.
My point of this, is that what librarians and higher education needs to develop are forums to share experiences, ideas, and experiments. Standards and policies will come, but it is just too soon. It is not too soon to share experiences and ideas.
This is where I sound like a broken record, this is what librarians should be doing in all areas of their work. The library should be a platform for people with common interests to come together to create knowledge. We already see a ton of librarians poking at these systems, and checking them for credibility, currency, and so on. We need to share that amongst the profession, sure, but we really need to be informing our constituents in their work about what we are finding.
• Where do you see academic libraries, especially university libraries, by 2030?
Well, if I do my job well, thriving. Thriving, but less driven by standards and best practices, and more driven by the unique nature of the universities they are serving. 50% of our budgets will shift from big deals with database vendors to open access deals with publishers. Our buildings will, by and large, continue to shift from quiet buildings with loud rooms, to loud buildings with quiet rooms.
We will continue to see our community of students and faculty as people, not transactions. We will work across the university on determining student outcomes and wellness.
* * * * *
Note • All the answers/ opinions expressed in this document are of the interviewee.
Cite • Hulagabali, Santosh C. (2023 June, 12). R David Lankes: Mission of libraries has to be improving society through facilitating knowledge creation in communities. [Blog post].Retrieved from: https://openinterview.org/2023/06/12/r-david-lankes-mission-of-libraries-has-to-be-improving-society-through-facilitating-knowledge-creation-in-communities/
Santosh C Hulagabali, PhD is an Editor of Open Interview. He heads Central Library and Publication Division of Central University of Haryana. He is passionate about anything that is creative, challenging and positively impacts self and others. Email: santosh@cuh.ac.in
I’m happy to read this interview by Prof.Santosh C. Hulagabali. (We were together at Karnatak University in 2002-04). Now I came to know a lot on role of libraries play in handling many issues of copyright, AI generative technology, services, etc.
Prof. R. David Lankes has shown a path to go ahead with AI [with its various manifestations like ChatGPT]. Dr. Santosh Hulagabali has aptly interviewed Prof. Lankes in light of future of academic librarianship.