 Thank you for the research libraries UK for having us today tell you a little bit more about our research and the workshops We've been doing and trying to get engaged with AI So we're going to tell you a little bit about engaging with artificial intelligence specifically in research libraries and university libraries So this is us as you've seen already And this is just to give you a sense of the research that we've conducted so far. So we have quite an expensive Research program And we've done several of these steps already. So we've started with an environmental scan, which Amanda will tell you more about To see what the state was in academic libraries right now. How prepared are they for AI? What kind of initiatives are they doing? And then we engaged in a survey of librarians perceptions of artificial intelligence How do librarians conceive of it? What do they think it is? The impact you can have on the profession and the different services that we offer. So this is kind of where we're at Next steps we want to Engage in and we've started doing already is some device testing. So testing virtual Assistants like a Google home or Siri on your phone The Amazon Alexa and seeing how well can they answer reference questions? So we've done a bit of that testing already But we'd like to kind of involve students and other librarians in that process. So because of COVID we Unfortunately cannot do that at the moment, but hopefully we will be able to engage in that soon And then we also want to survey Students perceptions of AI like how do our students think about it? And what impact do they think it will have for their research process and information seeking behaviors? And do some more device testing with the students. So we're sort of at the midway point of our plan and pending COVID restrictions being lifted will be able to continue engaging in these So I'll turn it over to Amanda who'll tell you a little bit more about what we've been doing Yeah, definitely So I think actually a huge part that we should probably purposes are really our fascination with virtual assistants Or voice assistants digital assistants. Have you're gonna call them? So one of the things that we were really interested in the way that you can use these devices to seek out information and the shift toward like voice queries, I guess and so we Wanted to see how well these devices were actually capable of answering reference questions That's a huge kind of part of where that device testing is and so one of the interactive things We're gonna do today is actually test some of that out of it So hopefully that will be a little bit fun But to provide a little bit of context We came up with our own working AI definition If you've been trying to see go a high definitions in the past And you know, this is a really convoluted and contentious in an area a lot of experts don't agree on what a definition of AI should be And so we like to actually think this is our living definition. We've added it I think like two or three times now where we've shifted around things taken out words added new concepts but our working definitions that we define AI The development of machines to accomplish tasks and reproduce the processes that are normally seen in humans And that the simulation of intelligent behavior is unique from other Automations as it requires the computer to use human reasoning or thinking to perform tasks And we like to call yet that the second sentence here is Aimed a little bit more at a more futuristic element of AI that we are we're slowly broaching this area of of human or computer reasoning and computer thinking To replicate on par a lot of of human tasks and so We're not always quite there yet But I think we can see in a lot of the technology that we do encounter that these computers That these technologies are advancing quite quickly And so one of the things that we like to show is our actual framework of AI and so Sandy will talk about this a little bit later when we talk about our workshops But when we do show our framework, we like to introduce AI as a family tree I know some of the analogies up there will compare AI to a person So you think that you know a person can speak they can talk they can see they can make decisions And those are all the different types of things that that AI technologies are trying to replicate And we kind of like to see it more branched out as a family tree So we have artificial intelligence joining with machine learning joining with data science to produce Many different subsets of AI and that's one of the important things that we really started to push is Having conversations about these subsets. So let's talk about predictive analytics and recommender systems in Libraries because that's a huge part of our journal databases now Or we can talk about neural networks and deep learning to be able to Catalog and classify information and to be able to part and make decisions about things that we might make choices on Where Sandy and I are mostly focused is within natural language processing Specifically within things like text generation and questioning answering We are that's that's usually where a lot of the virtual Assistance live but then also the tech that's powering that is coming from speech and vision or speech to text and text to speech So you can see how AI is not just one thing. We don't just have an AI We have all of these different technologies that make up a larger umbrella of AI and so What we want to kind of start doing is is having these smaller conversations about these different topics and And making people feel more comfortable with speaking about neural networks versus just AI or talking about machine learning or deep learning And not just yeah, like we said So really getting people comfortable with that and one of the ways that we do that is through our workshop series Which sandy will will break down in a little bit And so Just to kind of touch into the research One of the things that we did was our environmental scan. So to Kind of take a deep dive into how universities were planning for AI. We did a survey of North American libraries to see, you know, how we felt and We did this from July 21st to September 6th 2019. We had I think 163 full responses from academic librarians initially We tried to do the survey as wide as possible So we were looking for public librarians special government health like anything and everything we could get and we just wanted to see What responses would come back and in the end to make the the survey as As accurate as possible. It turned out to just be an academic library focus But we did get some really interesting information out of some of the other other data So hopefully somebody takes that up to look at how other librarians outside academia are interacting with AI and how they feel about it But our research questions what we really Like I said, what we wanted to know was how did librarians perceive artificial intelligence? Because you know, we're so caught up in that that larger broad term of AI We know that not a lot of librarians Even know that they're interacting with it sometimes. So we wanted to see how they were perceiving it We wanted to know if librarians were conscious of the fact that they were using AI technologies And we wanted to know if they had been providing any support around AI technologies whether through instruction resources that they were using Or or anything else if that may be in the back end And so we had some really interesting findings from that and this was in our our perceptions of artificial intelligence paper Which I put a citation in here if anyone wants to to read the full thing, but I'll give you some of the highlights Right now, so we found that 77 percent of librarians indicated personal use of AI at home and so this structures is like really really interesting because As a librarian, I don't think you can actually get away from AI It is in everything that we do inherently even if we don't know it I mean you could say that you interact with artificial intelligence just through doing a google search, you know You can interact with AI just from using the databases in your library So it's one of those things that really opens up the fact that You know at home and at work people might not even realize that they are interacting with artificial intelligence is is only 77 had indicated And then we also wanted to look at how librarians were interacting with virtual assistants and we found that only 60 indicated that they used them And that was really interesting to us Just to see that number kind of grow because we've noticed like the large trends and purchases of smart home devices in the use of Siri and I think it'll actually be interesting to see in kind of I don't know if post-covid is right word, but a post-covid world How interactions with these types of devices have grown You know people have had more time to sit at home and test things out and and kind of grow that algorithm inside Maybe I do want to play around with Siri a little bit or maybe they've gone in the complete Absolute opposite direction where they're now completely terrified of this technology and they don't want to tracking them They don't want to interact with it or feed it any data But at this point we found that about 60 percent of librarians were interacting with virtual assistants in one way or another And then we found that only eight percent believed that AI was being used in their library. So this is a really nice kind of comparison about 77 percent of librarians using AI at home So as we all know it AI is pretty entrenched in in our digital world. It's very hard to escape It's in Netflix. It's in Pinterest. It's in Amazon if you're making purchases there, like I said, it's in all almost all of our resources that are subscriptions that we subscribe to for for databases So to say that we're not Using AI in our libraries is kind of a huge eye-opener that only only eight percent believe that they were actually consciously using it So that told us we probably need to do a lot more outreach about artificial intelligence libraries because People might know that they're using recommender systems. They might know that they're using Computer vision to do digitization But they might not realize that that is actually a component of AI And so that kind of gave us a direction of where we should probably head with some of our our future planning And kind of striking me as well Only 20 percent of librarians believe that their patrons were interested in using AI And so this here is I think really going to set up a good kind of A good mirror to when we do our student perception survey to see how our patrons our students or community Wants to interact with AI if they have the desire to do that But the fact that that only 20 percent Believed that that people were interested See a little startlingly low to me, but at the same time Also quite high for the fact that We they're kind of that not a lot of people were actually doing AI programming So 20 percent is is you know, it's not it's not completely high, but it's not an insignificant number either That's you know one fifth of your your population is Potentially like interested in taking advantage of a survey and I think as librarians We can all agree that we've done way more for for way smaller groups of people So it seems to me that we should be jumping on this, but also that those numbers should be growing And of course Where we get into our virtual assistance area 36 percent Believe that AI could replace a librarian's job and there was actually a really unique study that came out in the UK actually that compared Like thousands of different professions and the likelihood that these would be We replaced by an believe librarians Felt somewhere around the 60 percent range that we had a 60 percent likelihood So it's interesting here that only 36 believe that their jobs would be replaced Now sandy and I I think are very much in the camp where we don't think our actual jobs are going to be replaced But rather maybe some facets of what we do could be supplemented with AI But that doesn't mean that we're necessarily going to lose our jobs It just means that we have to learn to coexist and interact with AI in different ways So it was it was good to see that you know people were a little bit more optimistic about about this question here But oddly enough 50 believe that AI could provide accurate reference assistance So this is a big part of what sandy and I are getting in the device testing Which kind of all started out with just asking Siri and Alexa and like google assistant Random reference questions to see if they could answer it and you know When we would kind of just ask the book casually as we started this project a lot of people like no No, siri can't answer something like that And you know, then we started testing and we're like, oh no, siri can't So can the google assistant it's it's it's crazy how Smart these devices are getting and it seems like you think we would know that ahead of time But so it seems like librarians are a little bit on the fence So perhaps half of us thinking that AI and Specifically some of those programs like the virtual assistants could Do reference work and then the other half quite not quite so sure that it would be up to par with like a reference reference interaction So that was uh, some of the more I think interesting findings of of our survey And we had some really great Questions that came out. I think we had asked people to provide us their own definition of AI to kind of gauge people's understanding We also had asked people about the programs that they ran in their libraries And there was a lot of questions about AI ownership as well So not a lot of people were sure whether or not Like their libraries could constitute using AI or owning AI Because it was part of something that they were subscribing to or something that like they used because it was free On another website and whether or not that belonged to the library I like to think that if you're out there if you're consciously using it If you understand how it's going that but you can say that you are actively using AI in your library That's that's part of the camp I belong to But this is kind of the core of our our survey here And the flip side of that is how libraries themselves are interacting with AI And so this is the environmental scan that we did Where we wanted to see if libraries were strategically planning for AI This is and this actually took place a little bit before our perception survey came out So it was it was interesting to see The results of the survey and then how they they matched up to the perceptions here And so this we actually did the summer of 2019. So before we released the survey We evaluated 27 research intensive universities in the u.s. In Canada And so we used the u15 in Canada, which is a group of 15 research institutions Sometimes known as like maybe like the top schools in Canada Because of that, but it also depends on on other rankings as well but it's a really good benchmark for trying to figure out who has research intensive programs And to kind of compare that we use the times higher education top 10 research institutions in the u.s And then we added in the University of Rhode Island and Oklahoma University during a second round of reviews Just because They had been pinpointed as heavy AI users in their libraries And we wanted to see if that made any kind of difference when we when we compared them to the top research universities And so we scoured every single strategic plan mission statement anything to do with emerging technologies To look for terms like artificial intelligence machine learning deep learning We wanted to find if there were AI hubs labs all that kind of stuff We looked through subject guides. We looked to see if there was workshops being offered on AI like were people explicitly providing programming Was there, you know tech that could be rented or borrowed or any that kind of stuff? And we also took a look to see the larger university presence So was the university offering, you know courses on AI did they have prominent researchers in the field that kind of thing just to make sure that there would be a connection And here you can see the list of schools that we Actually took a look at so our first two boxes here are the u-15 in Canada And then our top 10 in the u.s plus Rhode Island and Oklahoma And what we found was that no strategic plans actually mentioned AI So there were no no planning and no documentation around AI. We found that Additionally, there was no programming explicitly, I think there was about five I'm sorry, there the five programming, but no like events. No no workshops Things like that that that explicitly called it out, but we did find The samford university library had an AI digitization project going on The Rhode Island university libraries actually does have an AI lab that they had been building at that time But again, it's nowhere in their documentation. Do they provide goals or you know strategies for how this is going to work And so you can see that like some of the grassroots AI stuff is happening But it's not quite making its way into overarching like administration planning for libraries Which is where we think some of this should be coming And interesting to know that all all campuses did offer Heavy AI presence. They had distinguished scholars. They had programs hubs student groups for AI So there's a lot of demand for artificial intelligence like information on university campuses but Only five libraries that we found were even kind of touching on it And that was just in the small stuff that we looked at and we know that there's a lot more doing Kind of like they said those those grassroots efforts But the thing about that is that they don't make their way into, you know, the larger library ecosystem as quickly or as visibly And so they're a lot harder to find a lot harder to pick up people who are doing things Um, and that was a couple years ago. So we've definitely seen a huge boom in AI activities since then um, and actually one of um, one of my favorite ones that we found was the 99 AI challenge at the University of Toronto libraries Which actually launched I think it was like two months after we finished doing the scan. So it couldn't even be included in the project Um, because at the time they didn't have anything to offer and then a few months later after after the research was finished The University of Toronto came out with this huge program where they brought in 99 faculty students staff community members to do like a big Learning discussion forum throughout the year to discuss AI its impacts benefits potential fears policy changes all that kind of stuff So it was really cool to see this huge library event happen But it's just a shame that we it just missed the the frame of our methodology there And i'm actually going to flip gears and turn this back to sandy. So we kind of took a look at the state of AI In terms of libraries how they're interacting with it. We looked at patron perceptions and sandy's just going to touch on our our workshops Yeah, great. Um, thank you, Amanda. So this Research so both our environmental scan and perceptions from librarians Um, kind of got us thinking and specifically Amanda It came up with the idea that we should do a workshop to tell people a little bit more about AI So I want to preface this by saying that we are not computer scientists So we were not going to go into Like a back end black box. How does AI actually actually work on a computer science basis? We really wanted it to be open to everyone No matter the faculty no matter the background Um, anyone could come to the workshops learn more about it if they had an interest and participate In discussions. So if we go to the next slide Um, thank you. So it's actually a series of three workshops. Um, just because there's so much content We didn't want to just cram it into one. Um, so it's three. So the first workshop focuses on AI literacy So the terminology around AI that great family tree that we'll go back to in a minute Um, but really helping people gain a basic understanding of AI Because sometimes with sci-fi and movies people think of AI as You know the terminator or things taking over the world They don't necessarily think of it as oh, it could be part of netflix recommending new movies to me So kind of bringing it down to a more manageable level for people to understand how it's actually used in everyday life And also the impact it can have on library research And what we've done is created a tool to help assess information about AI So if you're familiar A few years ago librarians in the us created a tool called the crap test To help people evaluate information, right? So it's it's an acronym and stands for the various things that people should look for so we kind of took that Basic premise but turned it into something that would be more appropriate for AI specifically So when the people are reading information You know in the news or hearing about it finding an article What different aspects they should look for and what should they question as they are hearing about AI and learning about it so that was our first kind of Workshop to make sure that people were on the same level and understood what it means um, I should say too that People didn't have to take these in order Right if you were interested in ethics and bias You could come to the ethics and bias and we always had the beginning have a short introduction to that terminology To make sure that everyone's on the same page So the ethics and bias ones really looked more into the policies on AI personal information And it was really discussion and case study based And people were really Willing to engage so that's the thing we found with all of these workshops is that we'd booked I think the first one we did in person When we could do workshops in person We'd booked two hours But even with two hours like people were were talking in their groups at their tables They wanted to exchange on these topics So people are really engaged and really willing to participate which is great um So the ethics and bias one got a lot of really interesting discussions going It's also interesting that we got feedback. So some computer science students actually attended And they really mentioned to us that these issues are not discussed in their programs Like they learn how to code they learn to do computer science But they don't actually talk about the ethics and bias of it which we thought was quite interesting So it's clearly filling in need And the last workshop focused on AI in research. So not Research on AI, but how you can use AI in your research So if you want to use some AI tools or systems, it's a kind of augment your research processes So which tools are around and are already being used So we give some examples And also how you would want to include AI in the research process and how to acknowledge it So if you're going to create a research plan or a data management plan Similarly, how would you acknowledge the use of AI? within your research So we offered the first workshop in early march 2020. So we were just under the wire to do it in person And like I said, we had a lot of engagement. We had students from a lot of different faculties Undergrads graduate students We also had a lot of people from the miguel lifelong learning community So maybe very often like people who retired and decide to take classes To keep learning things. So we had a lot of people from that alumni like all the different programs at miguel And different types of members from the miguel community So it was really interesting to see all these people kind of come together to talk about this We did have to move it online because of the pandemic Which means we could yes reach maybe more people, but it's a little bit Different to try to have a live discussion online as opposed to in person So we'll see how that goes when we hopefully can go back in person But we got a lot of participation involvement. Like I said, we did Have 76 participants and given that it's the first time we were offering these We're pretty happy with it. But we had a lot of engagement So if we go to the next slide So this is again a great family tree that we present every workshop To make sure that everyone has a like base understanding of what the terms that we use are and what they mean And how they rely to see each other, right? And then so this is a look at the evaluation tool that I mentioned. So we called it the robot Because I thought it was quite fitting So it stands for reliability objective bias ownership and type. So How reliable is the information shared about the AI, right? Are you reading it? In a newspaper, is it coming from the company? Like how reliable can you can you actually trust what's being written about it, right? Because sometimes if we read maybe more popular articles about AI Very often they tend to be either Like very doom and gloom or this is going to save the world. So sometimes there's not necessarily a lot of gray area Um, and then objective. What's the goal of the actual AI? So what is it trying to do? Um, and what's the goal about share of sharing information about it? um What could create bias or potential ethical issues Within an AI and that's where like a lot of discussions came up with, you know, facial recognition Softwares and privacy like those sorts of things people were very engaged in um Who's the owner of the AI technology because that can really impact who gets to use it Who gets access it and again with facial recognition softwares like there were some companies who were making it available to say, um Police enforcement, but they were talking about making it available to the general public as well, which you know causes a lot of Concern. Um, so those kinds of issues we were we would discuss And then type so which subtype of AI is it? So that's where we go back to kind of the tree that we were looking at Um, and is it more theoretical or applied right because some articles will Describe things that could theoretically be done by um Like AI or all powerful AI systems, but we may not be there yet So that's important for people to be aware of the difference between the two um And we had some like really great conversations around all of those points But it's just to give people kind of markers of what they should look for and what they should question Instead of thinking of things as all all black or white right that there's um a gradient of information in this. Yeah um And so if you are interested in learning more about um different AI Kind of applications and academic libraries because we got really interested in this um, we Actually edited a book with acrl. So the association of college and research libraries Um, it is in production. So hopefully we'll be out at this point like um Our shameless plug and there is a chapter on the national archives of the uk so But there's actually some of the really great things that we talked about like the 99 a a challenge the university road island AI lab There's some really good some good stuff in here both back end and um, like user services that we're really excited to share with people Yeah, that brings us to our interactive portion Um of this thing. So before we get into some discussion questions I wanted to give everyone a chance to actually do a kind of modified version of a touring test So if you've never done a touring test before if you haven't heard of them, um, it is a method of Examining an interaction with a computer to try to determine if you are talking to a computer versus a human That's a very kind of watered-down Explanation of a touring test and so what we have here is actually a reference question That was asked to a media librarian on our virtual chat system Which we then asked to siri and we asked to google assistant So i'm going to present three different scenarios to you now And we're going to see if we can figure out which one was the librarian And we do have a menti meter poll for people to put their answer into So I think the link there is in the chat So you can pull up the link to the chat or the link to the poll and We'll go through these i'll give a little bit of time and if you need to like go back to see a slide just let me know But so each interaction is the same So this is scenario one and so our user is asking how to cite a website in chicago And the librarian Responds here is a link i found for citing a website in chicago with a link to purdue owl Now purdue owl in north american academic libraries is a pretty Pretty standard source for For information on citations. I don't know if it's the same in the uk So I just wanted to clarify that this is a pretty scholarly academic source that that would be provided So this is scenario one I'll just leave that up there for a second for people to To take a look at And here's scenario two so same question How do I cite a website in chicago and the librarian respond citations for a website in chicago include first name last name quotations title of the web page Publishing organization or name of website publication date and or access date if available comma url After it has been used once in your reference list It can be last name comma shortened title more information can be found here and this is a link to a university library website on citations So our third and final scenario The user asks how do I cite a website in chicago and the librarian responds? Here's a link to the chicago manual style you can access online And it's a link to the same purdue owl page So i'm just going to pause sharing my screen for a second just to make sure I don't reveal The answer as I go But you do have the mentimeter Pull in there. So i'm going to resume sharing. Oh, it looks like some people have already filled this out perfect, so Everyone should be able to see my mentimeter screen sanny. Can you confirm? Yes, perfect. Okay. So it seems like most people think action uh scenario two was with the librarian Does anybody need me to put the scenarios back up on the screen or you feel like you've got a good enough understanding to think which one might be which Okay. Well, would it shock you to know? That let me press enter here. It was actually scenario number three was the real librarian so Interestingly enough scenario one was with siri, which if you take a look at the uh, like the text of the answer You can kind of get that siri feel for it. But here is the link um the response scenario two was google assistant And I do want to kind of preface it when we asked it this question I used both my smart speaker device and the google assistant app on my phone And the google assistant phone app when it gives me the answer to this question It actually brings up a picture of how a chicago website is cited and step by step Will break down where each part of the citation goes in both the long form and the short form Before providing you a link to a library website And the third answer was actually from a librarian from our own virtual reference service um And so you can see that the they had kind of just given the ala link There was a longer conversation where they asked about other things that weren't related to citations But that was the citations part that I took out of that conversation where it was how do I cite this and That was the link sent so oddly enough It was uh It was number three was the real librarian and scenario two which most people thought was was the human was actually google assistant and uh when sandi and I Did our device testing we were like startled by how well the google assistant the google home could answer a lot of these reference questions um and As I see a chat there it's reassuring to know someone other than librarians are using them guys and that's actually part of the the results that we found was that google assistant is really good at offering up lib guides as like information access points when you ask it questions And part of what we want to know is that is that because it's us who's asking and you know We tend to search for lib guides or other people's library Pages or is this going to vary by person? And so when we do device testing and we want we want to include students we want to have you know Completely blank profiles. We want to have users who are heavily invested in their virtual devices who search You know pages we want to see how those answers vary depending on who's the person asking the question You know What their google algorithm or their series is is kind of set to to give them But it was it was impressive The the ability for for google assistant to answer Um, so thank you for participating in our our little test Hopefully now maybe you'll be one of those people who believes that virtual assistants do have the ability to answer reference questions Maybe that 50 that we looked at is is shifting No, we do have some discussion questions here. I'll let sandy take it away, but also I believe you can ask questions in the q&a part And if anyone wants to ask a question live You would just raise your hand and you'll be able to I think turn on your mic in your video and sandy And I are happy to answer but we do have questions for you in case Anyone needs a little nudge Um, thank you. Yeah, so these are essentially just um low questions to get you thinking so feel free to Discuss or ask questions about whatever you like Um, but we were curious to see what are some of your own perceptions of ai Um, is ai included in your strategic plan or maybe it will be now Um, or how could you perhaps include it? Um, which ai applications do you use at your library? Are they more behind the scenes for cataloging meta data? Maybe you're using linked data? Um, or is it more like the on the user side right like workshops or reference services chat bots or something I'm looking into as well. Um, and do you have any ai programming in your own library? Um, but if not feel free to ask us any questions I do see that there was a Q&A from earlier. Um, so could the ability of machines to make intelligent decisions be risky without having human emotion? Um, I would just say yes, I would flat out respond to yes, that would be very risky Sandy and I are really interested in human assisted ai Uh, specifically and I think that's a huge trend in a lot of ai research right now is having, you know, human But either explainable ai or human assisted ai so making sure that there is a human element in any decision making processes um, I actually attended an Open ai discussion held by the government of canada where they were pooling people to kind of see how they interacted Um with ai but also what we thought about policies And there were a lot of scenarios where they presented to us like if you know, ai was used to make decisions regarding immigration Or regarding any sort of government processes How would that work and what kind of human element would be involved? And it was it was really I guess positive to see that like that was in the forefront of discussion So like how do we integrate humans into each part of this? Which is why I don't think librarians are going to lose their jobs anytime soon. There's always going to be a human element to it Yeah, um, I think someone's asking in the chat too. Um, that if you're using those virtual assistants And it's a speaker and it reads the url to you. It's not very useful. Could you send the link? Actually, most of those so when you use a google home, for example, yes, I speak to my speaker But there's also an app and in the app on your phone. It would send you the information So very often it will send you like Amanda said that guide or Kind of like a screenshot or a link where you can see the answer In the guide. So there is a way for you to find it again or follow the link through the information Yeah, it's really good Yeah Just to jump in briefly I've seen there's been a few comments coming in that have gone only to panelists that may have been intended that way It's fine But I just wanted to say if you want to share something in the wider chat Make sure that you click the box for all panelists and attendees if you want everyone Everyone to see those um I have a question if I may and I've sort of been been reflecting on something that you said Or rather some of your server results where only very small number of librarians Said the library is using AI What I would find really interesting is is there any way to find out how close to the reality that actually is I mean, I know you've you've looked at the websites and so on but I think strategic plans may or may not tell you what systems are being used And the reason why I'm asking this is in a way So you've just asked the audience is the library using AI But interestingly if we don't know that systems that we have are using AI technologies How can we be aware of what the risks and benefits are and so Building on to this Is there perhaps a risk that we are not aware of how machines are making decisions already? And how can we deal with this? Yeah, so I mean to kind of jump to the last part of that. Yes, there is a huge risk that we're not Like not well enough aware and that's I think a huge part of why we're doing these AI literacy workshops is to start providing more opportunities to become aware and to become part of the conversation um I would say to them that my question would be to everybody if you're not sure whether or not AI is a part of your library Um, I would say do you have a digital scholarship? Presence at your your institution. Do you have any kind of maker space where you interact with with certain technologies? What are those technologies and start thinking about, you know Some of those like data visualization software and tools that you might recommend to users or that you use for your own work Consider consider those. Um, I think Sandy and I come from that user services perspective So we tend to think of those first. Um, I did see somebody mentioned in in one of those comments about using Um, uh, sort of scanning material for an OA repository for I think it was um Sorry, so to scan digital heritage create an OA picture repository Someone had sent to us and so for something like that There's a lot of a lot of AI applications going on in the background that like Sandy and I don't always get to interact with But that we know are happening at our library But maybe those librarians working with them don't understand that they're they're using artificial intelligence when they are interacting So that becomes I guess part of the conversation that we want to spark is is Who is I guess a conscious of what they're using and if they're not how can we ask them other questions? So that would be like the next thing that we would do is ask you about your maker spaces Ask you about the resources that you interact with what do you? Do behind the scenes which you promote for patrons and then we can start to figure out What of that is AI and how do we make ourselves more aware of it? um And I think like slightly different but someone was asking in the chat about the risks of Disinformation when using smart speakers to ask questions and do research And that's something I was really interested in because I'm a librarian for religious studies And we know that if we research anything related to religion online Depending on which religion it is You know, it could be prone to disinformation or you know, even slander in some cases and just false information So I was really interested in seeing how we would respond to some of those questions And that's why I think it's important like Amanda said to really engage in Kind of educating people and providing instruction about this because much like You know, even if you use scholarly research database, you could find predatory publishers You could find research that has not been vetted properly, right? So you still need to have an element of evaluation to it And it is much the same for smart speakers, right? It's one thing if I ask it Like I don't know how long do giraffes live like a basic fact factual question It's another one if I'm asking for information to create And build an opinion in a research paper So we really emphasize that evaluation element and like Amanda said, that's why I don't think we're going anywhere anytime soon because So much of our job is showing people how to find and evaluate information Yeah, we don't stop being information literate just because we search in a different way, I guess or that we interact in technology in a different way Thank you I think there's probably quite a few more questions I could ask about bias, but we have a few more Q&As from the audience So one Question is I'm interested in the role library librarians could play in building AI behind the scenes I understand that I learned from constantly asking questions to set context and build its own future connections Do you think librarians could have a role answering AI questions in the future? So I guess this is probably a question aimed at our role in training AI to make it more useful and help it learn better I think so personally and I think sanny would agree that you know, we have huge corpora of like data and and in Just like information sets that could be used to help train. I we have a lot of training data I'll say and so to become more involved in that I think would be really great for libraries I think there's maybe like risk aversion and hesitancy. There's also privacy considerations to be made When it comes to how we use that data, but I believe that librarians should be you know When they are able and if they're interested should be actively trying to participate in those conversations Yeah, and I think specifically for we've been doing a lot of research with voice assistants and and reference interactions specifically um Is a big area of research of mine as well and virtual reference And so far like those virtual assistants are great if we ask them a question Like how do I cite something in chicago? And it will give it to us But it's not going to engage in what we tend to call a reference interview like it's not going to you know Ask about is it for a class like how many sources do you need? What kind of sources do you need? Like it's not going to engage in that kind of more in-depth um querying Unless unless yeah Yes, that's what amand is working on but um, that is something that librarians could help inform right is that um providing that kind of knowledge Thank you. We've since I have two more questions And as we are sort of slowly coming to the end, I'll just maybe Read both of them and then maybe you can you can tackle both Together So the first one is in your review Did you come across any notable examples of AI being developed or used in conjunction with other professional services for research or student support? For example analytics and somewhat linked to this the other question is What would you recommend as AI services for librarians and users to experiment with and evaluate in an ag library context? so On the one hand What have you come across in other professional services and on the other hand? What do you think librarians should should keep an eye on? Well, I can answer the first one quickly because that has a I think pretty quick answer And then Sandy maybe you want to talk about some of the stuff happening in the digital scholarship hub so It was not really part of the scan to to look at some of those those back-end services that were happening So we didn't consciously pick those up I have you know read articles and seen instances of AI being used, you know in like grading type decisions or in Even sometimes like disciplinary type things to look at like plagiarism and cheating But it wasn't something that we were consciously looking to grab so I don't have a lot of information on that unfortunately Yeah And in terms of like AI services for librarians and users We've seen really a wide variety of What is possible to offer and I think It can be done at very low cost and it can be done like at very high cost like there's a You know, there's a wide variety of what's available. So what we focused on was really things that Would just take us our time for example, like we didn't have to pay for tech um, and then when you look at other institutions like Rhode Island that are engaging in a very great AI Hub where people can go and you know have Arduino kits and Raspberry Pies and like all these kinds of different technologies and High efficiency computing and all of these things so they can clearly maybe have a little bit more budget for that We went for the lower option Um, but some things we saw that were really interesting were really kind of grassroots initiatives where librarians with starts Journal clubs just because they wanted to learn more about the topic for example, and I think that's something we see a lot too is that Like specifically user services librarians Maybe don't have the knowledge or are not comfortable talking about these topics Like I know I wasn't because I don't have a background in computer science But a lot of people are kind of educating themselves with journal clubs having workshops We have a digital scholarship hub at Miguel now where we offer a lot of different programming So we offer our AI workshops through that, but we also have like digital scholarship data visualization um Three printing like all these kinds of things um through the digital scholarship hub um Yeah, so there's there's really a wide variety of what's possible to be done