Embedding Technology into the New Strategic Plan at the Library of Congress

Gerry Bayne: Welcome to the Coalition for Networked Information Podcast. I'm Gerry Bayne for EDUCAUSE Review, and this interview features a conversation with Judith Conklin, chief Information Officer for the Library of Congress. You can find more conversations from the CNI meetings by visiting the EDUCAUSE Review website at er.edu cause.edu. Here's our conversation.

Gerry Bayne: What are some of the digital strategies that are embedded in the new Library of Congress strategic plan?

Judith Conklin: So I would like to first start about the previous strategic plan for the Library of Congress. It was 2016 and it was a five-year plan, and we collaborated with all the business leaders, but we also at that time determined we should have a digital strategic plan. And so it was separate from the Library of Congress strategic plan, and we went for five years with that. Within the Library of Congress strategic plan, everything we did had to be pointed to our goals of the Library of Congress strategic plan. What we found out in that five years was the business leaders and the staff were not pointing to the digital strategic plan with their goals and their directional plans, what we call them, their business directional plans. They didn't necessarily look at that document. So when we, about a year before the new strategic plan, our strategic planning office brought the business leaders together in numerous sessions, and it was clear to us through that collaboration that digital needed to be baked into the strategic plan, the new strategic plan. And we, the Librarian of Congress, the Deputy Librarian of Congress and myself, made a decision to forego a separate digital strategic plan, but instead bake it in the strategic plan.

Gerry Bayne: Now, what was the reason you felt that that needed to be baked in rather than a separate plan pointing to the larger strategic plan? Did you feel like the separateness added some complexity that didn't need to be there?

Judith Conklin: Well, one, they didn't always look to the digital strategic plan. It was a separate document and sometimes people would forget it even existed with it baked in. I like to say that digital is baked in the new strategic plan, but more importantly, technology and digital is baked in the business. Now, if technology is not baked in the business, then a CIO is not doing it right. It needs to be baked in the business. We are no longer service providers only. We need to be baked in to the strategic plans into everything that we do because digital and technology is part of the world and it will never go away. Therefore, it needs to be in that thought process. So now that it's in the strategic plan, not only in that collaboration of developing the strategic plan and all the business leaders at the table and their people at the table, they thought very, very carefully about what do we need from a digital perspective as well as other things that should be in the new strategic plan.

Judith Conklin: So if you read our strategic plan, and it is on our public website, you can see the digital is absolutely throughout that strategic plan. The title for the strategic plan is A Library for All. When Dr. Hayden, the Librarian of Congress, became the new Librarian of Congress in September, 2016, she really wanted to open the treasure chest of the Library of Congress. And that had two aspects. The way I look at it, the first is physical, look at our reading rooms, 22 reading rooms, but also our exhibits, the Great Hall of the Jefferson building and that fantastic experience, but also she wanted the American public to see and experience our collections from a distance that they, as much as possible, they could see our collections and experience our collections without having to come to Capitol Hill and Washington DC. And so now we are doing more and more and more in a digital way to open up that treasure chest that the American public can experience us. And that's what her goal has always been since day one. And so I'm very excited about that. And so in the new strategic plan, you will see that what her thought was and what her directive to us was, so it's a fun time. It's wonderful to be at the Library of Congress. It's wonderful to be the Chief Information Officer at the Library of Congress. It's

Gerry Bayne: Exciting. What are you enjoying the most? Is there a specific aspect you can point to that you're like, ah, this is a lot of fun? This part?

Judith Conklin: What I enjoy the most is the being part of the business.

Judith Conklin: I'm not a service provider in their minds. Now I am a business partner, I'm a strategic partner, and we collaborate very much on what their needs are. And it's seeing the users. We have a term at the Library of Congress. It took us about a year to come up with it called users. And what does that mean? What does that word mean? It's the researchers people who want to register a copyright. It's Congress. Because we serve Congress, our mission is to serve Congress and the American people. And to see that treasure chest open and to be part of that, it's absolutely exciting to see all aspects of what we're doing for the institution and for Congress and for the American people. And just say, I had a part in that. People who work for me have a big part in that. And they're a fantastic group of people, and they love the mission. I love the mission.

Gerry Bayne: That's great. One of the things that is a huge issue in academia that we talk about a lot is what we call a seat at the table is CIOs not being the plumber, but being the actual architect in some ways to what we're doing. And that's sort of a newer

Judith Conklin: Trend,

Gerry Bayne: Very much in the past five to 10 years of CIOs having a seat at the table.

Judith Conklin: So let me talk about that.

Gerry Bayne: Sure, please do.

Judith Conklin: Back in 2000, there was no CIO, so the then Librarian of Congress, Dr. Billington had a study done, and in that study from National Science Foundation, and in that study it said, you need a CIO. And so we in September, 2000 did get one, and through the years they still saw the IT department, and it was kind of called that as that plumber as its service provider. And as late as 2016, that was the thought. Dr. Hayden is a rockstar, a fantastic leader. She came in and within the first month, she directed two things. There were issues with technology at the Library of Congress at the time, and I won't get into that, but it was a lot.

Judith Conklin: The first decision she made was to move the OCIO, the Office O, the Chief information Officer out from under the Chief operating Officer where all the services were and put this OCIO directly under her, meaning the CIO would work for her, not the number two, not the Deputy Librarian of Congress, but her. So I am fortunate enough not only to have a seat at the table, a board member, and in the Library of Congress, we called the executive committee, but I also worked directly for the Librarian of Congress. And that really makes a statement. She's very smart. That makes a statement to the other businesses of the library. She was saying, this needs to be strategic. They are not just a service provider. They need to be, and she doesn't use the term baked in, that's my phrase. But in essence, that's what she was saying. And she wants the business leaders, that executive committee, that board, she wants them to collaborate, to work professionally together to reach our end goals. And that is what we all do. And I am absolutely on the board.

Gerry Bayne: Is there anything else you'd like to say about the strategic plan?

Judith Conklin: We have numerous missions at the Library of Congress. I mentioned our mission statement, but we're different from many libraries in that the United States Copyright Office is in the Library of Congress, so that's a different kind of mission. The Congressional Research Service is they have analysts, 500 analysts that write reports and memos and advise Congress on many, many, many different issues. And so that's a different type of mission. And then we have the typical, what you see libraries, the catalogers and the librarians, the reference librarians, acquisition specialists for collections. And so that is baked in, excuse me, all of that you'll see in the strategic plan. But what do they need from a digital perspective that's in the strategic plan. So it shows how digital and technology or whatever word you want to use is helping or playing a part in the business side. And so it was very thoughtfully put together. The strategic plan took about a year.

Gerry Bayne: Wow. Okay. Great. In a description of your approach, it was said that the Library of Congress wanted to balance caution with the digital experimentation. What did you avoid in this plan in regards to caution

Judith Conklin: With digital experiments? First, I'd like to say that machine learning, AI, whatever term you want to use, it's a fast moving train. Anytime a technology is that fast, I think anyone, any organization should be aware and tread lightly and do it methodically. We do have, within OCIO, my business unit, we have a Digital Strategy Directorate. Within there, we have what we call LC Labs or LC Innovation, our innovation center. And we do AI experiments, machine learning experiments. We are not new to the game. We have been doing these type of experiments for about five years. And I really thought about that. I was at a conference in October where there were a lot of CIOs there, and they wanted to put a panel together about where CIOs were AI. And there were about 14 CIOs there. And I was near the end, which was very good.

Judith Conklin: And what I got out of the first 12 or 13 was they were just dipping their toe in the water. They were just getting started and they were still figuring it out. And that wasn't my story. My story was we've been at this and we've learned a lot, and we are doing it in a very strategic, a very methodical way. And so I thought about it, why are we ahead of the curve a bit on AI? And I got it down to this, and it's called LAMS Library archives and museums that people in the LAM world have a tendency to be just a little bit ahead, maybe a lot ahead than some of the other entities in the world in my world, federal agencies. And I got it down to this. Researchers, we at the Library of Congress have a tremendous amount of historic collections that are unbelievably useful not only to researchers, but to authors that write historical books. They come to our manuscripts reading rooms, often we know them, big name authors that want to use our collections. And so what can those researchers do with our digital collections? And obviously we do not have 100% of our collections digitized. And I'm asked that question a lot,

Gerry Bayne: I bet.

Judith Conklin: Are you done or when are you going to be done? And we're never going to be done because of the large amount of material collections material we have. But with the digitized material that we have, researchers want to do new and innovative things with the collections. So we have 22 reading rooms. The researchers know manuscripts. They know what to expect out of manuscripts. They'll go to prints and photograph to go see the same subject over in prints and photographs, and it goes from there. And all our geography and maps, well, with digitized collections and ai, wouldn't it be cool if they could use AI to span across our fantastic, rich historic data on a specific subject? And I think that's what AI is going to bring. But we are treading very lightly. So first of all, the AI tools out there, the AI models, in the five years we've been doing experimentation and we have not seen high percentage results.

Judith Conklin: And so a lot of it is about learning and the AI tools need to learn, but also why don't industry AI tools work well with our historic collections? Well, first of all, the AI tools and industry have a tendency to use the more recent English language, right? Well, our historic collections, let's say 1700s, 1800, AI tools don't do so well in that. And so we've had to, I don't want to say slow down, but very methodically determine how do we get around that? How do we train an AI tool with our historic type collections of the old language or even multiple language? We have over 400 collections and in over 400 languages. So how do we get around that? So we do it very methodically. The other is we are pushed a lot by different entities to use AI on business processes, on workflows to speed up some processes. And I will give you an example.

Judith Conklin: People may not realize the congress.gov is developed and maintained by the Library of Congress. In fact, the developers, software developers reside within OCIO, within my organization. And so congress.gov is the authoritative legislative information for the United States. It has all the bills and it tracks as a bill goes through its process. If it reaches the floor, if it becomes a law, the Congressional Research Service has experts and they have legislative analysts. Well, if you go to congress.gov, you'll see a bill summary, so you don't have to read the 200 page bill. There's a bill summary.

Judith Conklin: The Library of Congress and CRS do not have 100% of bill summaries completed. We never have. So we can call that a backlog. So Congress would like us to reduce that backlog and they would like us to use AI to do that. That's a very dangerous proposition for us to use AI because our AI experiments have shown that the percentage of success we'll use that term of accuracy, is low on our previous experiments. We do have a bill summary experiment going on right now. And so what we require out of the bill summaries is the authoritativeness. The nonpartisanship, it needs to be a nonpartisan bill summary. It can't lean in any direction, and it has to be trustworthy.

Judith Conklin: Another entity outside of the federal government took, because we have an API took bills, we allow this from congress.gov and did an experiment. They just put it through their AI tool, an AI tool, and they posted it and they put a disclaimer on what they posted, that these are bill summaries of the United States bills at this time of the current Congress, and it was done in AI, and it hasn't even been looked at by humans. The CRS analysts, legislative analysts took a couple of those and analyzed because that's what they do. Those AI bill summaries interesting. And it was atrocious.

Judith Conklin: It was not accurate. We've been asked, would you be willing to speed up, reduce that backlog, just run it through the AI tool, and then put a disclaimer. If you come to the Library of Congress website, loc.gov or congress.gov or copyright.gov, the American public researchers, other libraries, they expect it to be authoritative. So we cannot put a disclaimer on, that's a very specific example of something we will not do. Put a disclaimer, run an AI tool, our data through an AI tool, and then just say this because people don't read the disclaimers. Not everybody pays attention to the disclaimers, but more importantly, it's expected to be authoritative and nonpartisan. And so we will not do that. That's a very good example.

Gerry Bayne: Absolutely. That's like saying, well, homework's halfway done here. It's okay. This is kind of what you talked about, so I'm not sure what you want to say on this. What were some of the biggest challenges arriving at your plan and what were the greatest opportunities?

Judith Conklin: I talked about the strategic planning process, and that involved everybody. So collaboration is vital, having all parts of the business. And then within, as we incorporate digital, we have many, many other processes, software approval process. We have a agency data management initiative that what do we do with our data? Where is our data? What types of data? And so it's very important to us that we're process driven and we collaborate in our processes. For instance, the A DMI, we have a collaboration. It's agency wide, institution wide. We also have an AI working group that's institution wide. We have business partners there at the table. And so that's the process. The processes, we're very process driven

Judith Conklin: Because we're a federal government, we have to be process driven, but also it ensures we are using taxpayer money appropriately. And the biggest challenges, I mentioned before that AI and technology in general is a fast moving train. And so we have to be very careful on what we push out, but that makes it harder that it's so fast and our customers and our users want the new technology. And so that's a challenge. My challenge is I see the big picture. I see all these things we can do, and I want to quote Dr. Seuss and say, oh, the things we could do, and I'm just so excited on everything we can do, but we have to do it methodically and safely and deliberatively and strategic and collaboratively, a lot of buzzwords.

Gerry Bayne: That's so interesting because the technology's moving so fast.

Judith Conklin: Yes,

Gerry Bayne: But you have to move so slow.

Judith Conklin: I'm hoping it's not too slow.

Gerry Bayne: No, I don't mean that in a critical way. I'm just saying you see all the things you'd like to do,

Judith Conklin: But you

Gerry Bayne: Have to move slow very, because you've got to be careful. So that's got to be a little bit of attention. And

Judith Conklin: The last biggest challenge is as technology and digital continues to move in a fast way, we need to ensure that we do it safely. And what do I mean by safely? And that's the cybersecurity hitch. We need to ensure we don't open doors or that we continue to secure in a very methodical way and protect our data and protect our applications and our IT systems because we are a target, because we are federal government, we have congress.gov, we have the Copyright Office. We know we're a target. And so we're constantly changing that IT stack upgrading that IT stack. And with AI, we need to add tools and processes from a cybersecurity perspective to protect our users, to protect our historic data. And I feel that's a very big challenge. And that does, I don't want to say, I will say that it keeps me awake at night, not because we're not secure, but because the bad guys are staying awake at night.

Gerry Bayne: It's an arms race.

Judith Conklin: Yes, it's an arms race. That's exactly it. Who's moving faster? And we cannot say stagnant on our cybersecurity posture.

Gerry Bayne: Yeah, that's paramount, I'm sure. Yes. Is there anything you'd like the higher ed technology community to know about where the Library of Congress is headed?

Judith Conklin: There's a lot of exciting things, and I already talked about A Library for All, so I won't repeat that. But we have three pillars in our plan. Accessibility, discoverability, innovation, and Dr. Hayden's is that all statement is all Americans connect to the Library of Congress. We have APIs on losc.gov, and we continue to put more and more digitized content. So let me say it another way. We continue to digitize, so more collections, more historical collections,

Gerry Bayne: And that makes it more accessible as well.

Judith Conklin: And then more accessible to people don't have to fly in. And then we have the APIs. So higher ed education can use that for their curriculum and for their libraries. They can say, well, the Library of Congress has that. We don't have that in our collections, but the Library of Congress has that. That's exciting to me. There's APIs when we have that on congress.gov. So if it's higher education for lawyers students that they can use that. Then we also are moving from our existing ILS to what we're calling LCAP, but it's a Folio open source solution. And that will have APIs in that project. We are moving from, in the library world called MARC Records, and that's the MARC Records is a cataloging standard. We're moving to BIBFRAME and when we move to Folio to LCAP, but it's called Folio. There will be APIs on that.

Judith Conklin: And I think higher education will be interested in that. They'll be interested that we're moving to BIBFRAME from MARC standards, and they'll be interested that there'll be an API there. Currently we do have bulk download opportunities for them, but not through an API. So I think that will help. And then we just received our budget and in there Congress would like us to do an experiment on something called DNA storage. Interesting. And that's taking some of our historic digital collections and putting it in DNA. The experiment will be first doing that and then see if we can open it and read it. And it goes

Gerry Bayne: When you say, can you break that down for me a little bit? I'm not quite grasping.

Judith Conklin: Right. And it's not human DNA. It's actually DNA. And this is a new

Gerry Bayne: From the tree of the paper.

Judith Conklin: Yeah. Well, and of the actual collection information, the digitized information will be there. And the thought is 300 years from now, whereas we have 175 petabytes of storage right now on a certain kind of storage. Well, if you can imagine how hard it is, as fast as technology's moving, I have to then say, okay, now we've got to move from that storage to another storage because it changed. That's huge. It's huge. Well, the thought is if we put our digital collections or digitized collections in this new technology and Congress wants us to dip our toe into it, and they funded it for an experiment to reduce it down to our digitized collections, reduce it

Gerry Bayne: That small,

Judith Conklin: That small, and I keep going like this, like a little vial

Gerry Bayne: If you're listening folks, she's holding her hand about fingers, about an inch apart,

Judith Conklin: An inch and a half, and think of a little vial. And so what our idea is, and we were talking to Congress when it first came up and before it went into the bill and that the America 250 is coming, and that's the anniversary of America. And we are on the America 250 program, meaning project. And so our idea is, and I don't know if we'll follow if this will work, but our idea is, the reason I keep going with a little vial is I'm calling that we are calling that a time capsule, and we're going to put 10 of our most historic digitized collections like the draft Declaration of Independence in there and some other very historic digitized content of the Library of Congress. And we're going to put it in a time capsule of America 250.

Gerry Bayne: That's so cool.

Judith Conklin: That's very cool.

Gerry Bayne: Fun.

Judith Conklin: That's so

Gerry Bayne: That’s so much fun. That is great. This has been such a fascinating conversation, and I really appreciate your time, Judith, and your expertise and insight into what goes into the Library of Congress

Judith Conklin: Technology. Yes.

Gerry Bayne: Thank

Judith Conklin: You so much. Thank you. Thank you for inviting me.

Gerry Bayne: Absolutely. That was Judith Conklin, Chief Information Officer for the Library of Congress. I'm Gerry Bayne for EDUCAUSE. Thanks for listening. I.