 I was really excited when I saw the call for paper and there was a line in there about research libraries and researchers in 2030 because I thought that would be a really good opportunity to go out and publicly embarrass myself. And I expect people to come back to me in about 11 years and say it was all wrong. When I started on this little thought experiment, normally I approach these things by sort of my natural tendency trying to offer solutions and fix problems. And in this case it turned all rather a bit more gloomy and more problems than solutions. And then I decided to take a slightly more existentialist turn and go with black and white layout and present these as sort of slightly depressing personal thoughts. And to balance that I had very carefully being driven very carefully planned a joke which unfortunately Tim Hitchcock has stolen from me. Because by training I'm an early modern historian like he is and obviously he spoke about that he feels he's perhaps not that well set up to comment on libraries. It turns out actually he was quite well set up to do this. For me it's even more difficult because at best historians are good at explaining the present and I now try myself at a bit of prophecy. So I think the key thing if it becomes too gloomy or slightly too obscure to keep in mind is Torsten's not a prophet. Most likely quite a few things I'll get wrong but my main intention was to spark a little bit of a debate and share some of the concerns that I have personally. Speaking of sort of prophecy and predictions in general I thought well before I look ahead about 10 or 12 years let's look back and see what has happened about 10-12 years ago that at least some people didn't see coming and that really changed the world and one obvious thing to look at is the iPhone. That's arguably changed not just technology but how we as humans interact. And if you look at what some knowledgeable people said in 2007 there were a few Apple competitors but also experts in the industry who basically said this is never going to get any market share it's never going to go anywhere. Which I think is just highlighting we aren't particularly good at predicting the future and I want to claim I am. The approach that I've taken in the following is a relatively simple method for a thought experiment. Take a few things that we already not just see on the horizon but that are almost about ready to hit us into the face. And just assume take one route how these things could be fast forwarded about 10 years and then think what might the implications going to be. So I'm not pretending to give you any unknown unknowns to surprise you with some of that. It's really just mixing a few things together and then have a slightly depressing outcome after that. At least in parts I don't think it's all. Starting on a positive thing looking back about 10 years looking forward open access has come a long way. Let's just assume this has all worked out. And arguably I mean depending on what statistics we believe we're already comfortably about 50% open access. We now have plan S and other initiatives. It doesn't seem too unlikely to assume given another 11, 10 years that will broadly be there. We might not have maybe fixed some challenging issues like monographs completely. But in many areas I'm hoping we'll be able to say we've won the fight it's not all open. Surely that's a good thing. Which it is but it might have some interesting consequences for libraries regarding to what's our role. And I mean in a classical way all of libraries are very simplified as to collect as much content as we can. Sometimes stealing often buying and then make that discoverable and accessible and help people create knowledge from it. If you look right now where people discover knowledge is not normally our catalogue. It will be Google or it's specialized services like on paywall or the open access button and others. And if the knowledge is all out there the question is will we be good enough to have the kind of systems to help people discover this better than organizations like Google or others can. I mean maybe there might be some specialized solutions that we can license but we are ready at the point where the default discovery mechanisms that people use are free at the point of use. Broadly at least in the open space. And if that's not going to change then for finding content and accessing content people might not necessarily go to libraries anymore. At least not for that type of scholarly content. Now you could say that's okay. We just shift and we're already starting to do this. We help people to make their research openly available. We pay APCs, we run repositories and that's just sort of changing the model somewhat. We're not on the access side. We're more on the production side. We can still do something useful. And that's indeed what many research libraries are doing at the moment. But you could see this going various ways. For example just looking at Welcome. They've recently launched their own open research platform. If I was a slightly authoritarian research funder I might at some point say this is all too complicated. I'm paying for this platform. It's good enough. I don't care what my researchers say. At that point library support for this will become very minimal. Now you might say research funders aren't that authoritarian. This is maybe just my wishful thinking that they were in this context. We're doing lots of other things. We will still be able to support researchers by paying article processing charges and processing those. But if you look at how this is run at the moment, this is a very arduous complicated process that's way too labour intensive. And you can imagine various ways, even with today's technologies, to largely automate the whole payment process. There's probably a topic for a separate task but I might have some ideas on how we could do it. Even if we don't have it, I think my assumption is all these workflows over 10 more years will have to get a lot more efficient. There will be a lot more automatics. Even if we have still a role in there, it's quite likely to be a lot less than what we currently assume. A few years ago, when I started at Imperial College, there was no full-time post completely dedicated to Open Access. The Open Access team, when I left about two and a half years ago, had six full-time staff. Will we need that many people to process things that we could largely automate? I'm not sure we necessarily will. I don't want to say that looking after content and content development, collection development will go away completely. There will still be things that will cost something. Electronic resources, some databases that are important for research. Learning materials, some of those are fiendishly expensive. And it seems unlikely to me that they will all become open. At least I can't currently see a business model. So if you are concerned, I'm not saying that collection development for contemporary content will go away. But it might shrink significantly and be focused on some of those resources. Well, then we might say there are other useful things that we do such as work around persistent identifiers, libraries work with orchids, data sites, DUIs. We're looking at other types of identifiers. There surely is an important role for libraries. And there is at the moment. But if you sort of fast forward and say we have a few identifiers established that we currently don't have, this itself will and we want this to be a much more automated process where identifiers will be added to research objects to data at the source. This might be in the lab. This might be at a microscope. This might be at least at the point when someone is publishing something. And often, and I probably still will in the future tell the publisher before they tell their organisation. If at that point you add lots of identifiers about an object to it, there will be relatively little effort required in manual curation. That's really how we want it to be, but it also means that our role in curating this content will become somewhat diminished. And more interestingly, currently many libraries look after something like an institutional bibliography. They provide data for the current research information system. Some even manage that for their library. How valuable is it going to be to have a local system that points at a global graph of persistent identifiers when all the information that you want to track is somewhere out there in the cloud? I mean you can have your local data with the identifiers, but if you want to do some analysis, you have to follow these identifiers to where they link. And if you're doing this anyway, then yes, as a library, we might keep a local copy of the data that we consider relevant, but we probably couldn't keep all of the data. So you would probably, even if you are in the research office or somewhere, you'd probably work on data that is curated on a global level by lots of players where the data is generated and that isn't necessarily always in the library. So this would be a good thing, but it might not be as much of a role for libraries as we might currently think by looking at this. Now surely we've heard a lot about open science, more broadly research data management. This will be an important role for libraries. And currently it increasingly is, but I would also argue we are still in a slightly experimental space and phase of doing this. We are in a space where you can still locally tinker with a bit of software and set something up, but there are more commercial players getting into this market. We're starting to see some more mature workflows emerge. My assumption is that there will be more and more platforms for supporting open research, either managed in disciplines by large organisations like CERN or will be provided commercially. And the ability of libraries to develop something that will be as cutting edge and is nice to users does, I'm not sure it will be that big because that's quite costly. So if we think we will still run systems in 12 years that support open science, open research more broadly, I'm not sure how far that's the case. It doesn't mean to say that we'll not have a role in this at all, but it might not necessarily be running the university data repository because there might well be other organisations who can do this a lot better than we in our role, maybe much more of an interface to things that live out there somewhere in the cloud. And arguably that's in some ways how we have to do this. I would think our number one user interface in the future is not going to be to humans. It's going to be to software because humans will consume information that we give them through software. You can already see this at home potentially. If you told me 15 years ago, I would talk to a loudspeaker and it would be the most intelligent entity in the room. That would have really alarmed me, but I'm now somewhat more used to it. Obviously it's not the loudspeaker that does anything. There is some machine learning emerging artificial intelligence somewhere with Amazon sitting behind it. That's already giving me a lot of information. The same is happening already with Google. A simple search box is a main mechanism for which we interact with the world today. What we need to do arguably is getting ready of feeding all our data into software so then comes to humans, which brings up questions will we still need a library catalogue if people use other discovery tools to do this? Yes, we will have to have some information store that pushes the data out. But it might not even have to push it out because most likely that will be a system managed by another organisation living somewhere out in the cloud because we might not have the resource. In fact, even some current library management system providers are starting to encourage us to move later to the cloud where we might perhaps have a local backup copy but become increasingly dependent on the software of these organisations. Behind all of this will be machine learning and artificial intelligence algorithms which we might struggle to understand quite frankly. We already see some of these big corporate conglomerates. Facebook is training a lot of software, making a lot of money by working on the data that we give it for free every day. It seems unlikely that we'll be able to at least on our own as libraries to compete with this. We might be in a situation where all our data much more than we think now lives in some kind of cloud environment and is controlled by algorithms of an organisation where we don't have the insight and even where it is open source we might not have the skill to fully understand what's happening with our data. I think this is partly what Tim spoke about this morning looking at the changes that he sees in how libraries process information. It's an interesting challenge and I think it's one that we perhaps aren't quite as well set up in terms of the staff skills that we have available. As I indicated before we see an increasing concentration in the technology market. There are different predictions about at what time there will only be this many global big technology providers. I don't think the exact data is important but if you look right now about the market power that Google, Amazon and others have that comes from an enormous amount of computing power of data they can train their systems on of economies of scale and being able just to completely dominate a whole market it's not just having a few new solutions they increasingly move towards offering solutions for a whole market. That again, the idea that we'll be able as individual institutions to stand against is quite unlikely even when looking at open source. There may well be something where we can join our forces on libraries and say maybe we need to have some kind of joint investment in infrastructure or at least some trusted partnerships and principles how we want these organisations to work but unless we change the way how we work and just doing everything on our own running our own systems they will just walk all over us because they have better user experience they have better design and frankly more intelligent solutions than we can offer them. Interestingly, if you look at what some of the big publishers are doing Elsevier for example on his website is not describing themselves as publisher anymore they are a data analytics company and I think that's the direction that they are going so one prediction I would like to make is it's quite likely that the organisations that we currently pay a lot of money for to subscriptions in an open science world will pay a lot of money for their analytics services not because we wouldn't necessarily have a choice but I think in an open world you would have a choice who does your analytics if all the data is open but they have built up a lot of expertise and capacity to do this and it would be interesting to see who will be able to compete with them so it's quite likely that we'll deal with the same organisations just in a different environment and in that case I think there would definitely be a role for libraries and this is something that we are already doing I've just sort of picked two examples in the British Library about being a neutral and trusted forum to discuss some of these issues around information raising questions as we've done this for BL on data and the impact we are now going to be partner in a CDT on sort of safe artificial intelligence I think that's the key role that libraries would have and it's also I think still a role that we'll have when it comes to raising concerns about legal issues and trying to influence governments and the like or a lot of this will depend how much expertise we have on the topics I mentioned previously because we will need that expertise to understand complex digital systems to argue the needs of our users and to make our points as information professionals in them Now some of you might say that's fine, I have special collections and people will come for those so I don't need to be too concerned and I think to an extent that's true but if we think ahead let's assume you have special collections that you fully digitise and you do the right thing and you make them available as public domain CC0 or even CC by it's all out there, you've succeeded everyone in the world who has internet access can get to them will they get it through our systems or will they go through some software providers who will have really clever tools and will be able to suck our content in I mean this is still a success people will get this data, the information that will work with it and our main role that might be one of being a preservation provider because access happens through the systems of others if our preservation services are not run by us but by commercial companies then arguably the main thing that we at that point then contribute is paying a commercial company to preserve content of other people access through services of other commercial companies still good not to lose the content but perhaps not the first thought we have about curating our special collections Now arguably in some areas where we probably fail user expectations of special collections this is where we will still attract them which is where we haven't digitised material and that will still be quite a lot or where we have digitised it but where there are access restrictions and we can't easily make it available online in a way that's to reuse people will still come for this but arguably they come in the area where we fail the user expectation of having the material available preferably in both physical and digital form but certainly in digital form One area that's been mentioned where I'm convinced we're not failing our users and where we have lots of expertise and that we can build on is workspaces because it seems quite likely humans will still have bodies in 10 or 11 years I'm not making more long term predictions here so they want to sit somewhere and they want to work and I think libraries have been historically good and are still good I mean digital contents out there some people have predicted years ago and as you all know that's not what's happened in fact in some cases we've seen many more people come in but we have to keep evolving how we make space available and I think some of the ideas that we heard today about interdisciplinary spaces and facilitating some of these discussions will be useful but also I think opening up our libraries as much as we can to local communities this is anyway something the government would like to see it's also something that universities would like to see and generally I would argue healthy thing to do for research libraries but I would assume we'll get more push to do this but also that universities will encourage us to do more of this and we're certainly looking at this from the perspective of the British library I think we will have a role to play around these spaces in the particular atmosphere and being a neutral place of learning and studying and I'm pretty confident this will still attract people in 2023 I also think that we will have a role to play broadly around digital literacy skills and for those of us who in particular also look at audiences that come from the wider public inclusion and bringing people into digital environment we've heard quite a bit about this in talks previously so I don't want to say too much about it because there are some really good examples but I think to highlight in there will only be able to in 2023 help people to work if digital technology and information if our staff understand digital technology and information well enough so the point is to be made about GIS but it goes to other areas and libraries working on this transformation but it's challenging, it's not easy and it's really difficult in some areas looking at the salaries that we can pay and what people who have these skills can get in other areas is not easy even if we make the decision politically we are going to restructure, we are investing more we are starting to hire data scientists even that's not an easy transition to make so this was a quick run down through a few thoughts you can have very different perspectives on this but in summary I would assume some parts of the traditional role of libraries in supporting not just managing collections but also discovering access will diminish how far they will do this depends on partly how many commercial products will be so good that our users still want to have them and will still in the case of our libraries or funders be willing for us to spend money I would argue we will do even more of what we are already doing quite a lot of, we are just procuring services but we have to probably do this in a slightly different way because it's not just individual services if your whole data everything that you do with it is very how it's analysed isn't the cloud that's provided by someone else you want to make them sure you have people who understand how to do procurement who can analyse these systems who can understand what the consequences are of the choices that we make when we pick suppliers and if we don't do this there's a risk that will fail our users and arguably we will deal with a few very clever providers some non-commercial we might decide that we want to see more non-commercial suppliers but I think this will only happen in some areas if we as libraries get together and say either we support some of those or we may be set up some new ones we also I think will and I'm still convinced this have a role in supporting open science but it will be a lot less on supporting systems and it will be a lot more on supporting skills overall and I would also assume that we probably see the size of staffing shrink through budget pressures but also because being able to automate a lot more the interesting question I think in discussions of our funders will be how much of that money can we retain and invest into new services and how much will just disappear through ongoing push for efficiencies either way I think skills is a key part that we have to invest because ultimately what I'm convinced that having approached this journey from a technology point of view is that a lot of this will be about human relationships and how we work with members of the public but also the researchers and others to do what we can to enable them to navigate a digital space even though maybe our role is content provider and sometimes also a service provider will be somewhat diminished and on that depressing note or maybe hopeful note I'll finish