 Right, good morning everybody. Good morning. So my name is Chris Morrison and my name is Jane Sepper and we are the co-chairs of the Copyright and Online Learning Special Interest Group, a special interest group of ALT, the Association for Learning Technology. Yes and you're joining us for our 63rd webinar where we're going to be talking about the very hot topic of copyright and artificial intelligence. Hot topic, we do indeed have an expert special guest with us today. So, let us have a look at the running order, if you can get there. You're controlling the mouse this morning. I'm doing well, I'm doing well, I'm all over it. So, we don't have that many items of news. We haven't actually done a webinar though, for the summer, so we're back from our summer break. Although, in the summer we've still got that holiday to come. Absolutely. Unless they don't have to worry about school holidays, taking a holiday in September. Exactly, exactly. So, yeah, but we're going to just do a bit of copyright news and then we're really delighted. We've got Alex Venlon from University of Birmingham joining us to talk about copyright and AI. And then also a bit about what we've got coming up next after that, but we're hoping to have lots of time for discussion about this topic. We know it's come up on this copy seek, and so I'm sure that lots of you are ready and waiting with all sorts of questions to ask Alex. And yeah, we're looking forward to that. So, what have we been up to since we last met Chris? What's going on here? So, yeah, this was, I mean, as you know, Jane and I are consummate musicians. Professional musicians really. But what we were doing here, absolutely. What we were doing here is we were at the playful learning conference in Leicester. This was at the beginning of July. And we thought we would do a songwriting workshop. So this was us demonstrating one of our copyright jingles and how we came up with it. But then we asked other people to come up with the conference theme tune, either the conference theme tune or a song about their own practice area of expertise. And it was fantastic. It was brilliant, wasn't it? We had three teams. They actually all came up with the conference theme tune idea. And there is a podcast. There is. The Pedagodzilla podcast. Yes, yes, yes. I will go and find that link and I will put it in the chat. Yeah, yeah. So our friends who run a podcast called Pedagodzilla, which is all about teaching and learning with a kind of pop culture twist. I think it's a pop culture corp, they say. Yeah, so they recorded live the performances from the four bands. I think they were all put together. Yeah. And yeah, there's some great anecdotes from the playful learning conference. So it's big inspiration for us. So we have been out on the road a bit since we last met, haven't we? I've got the link here. Excellent. Good stuff. Can you even put that in the chat? Thank you. There we go. Right. So this is just a quick reminder to everybody about the webinar and archive of all the previous sessions that we've done. So you can find those on our website on Copyright Literacy. But you can also subscribe to the YouTube channel that comes from the Association for Learning Technology. They have a playlist that's got all our recorded webinars. This is where the recording on today's session will end up as well. So just a quick reminder, many regulars will know this. Okay, Chris. Play that theme tune. It seems buffering, buffering, loading, loading. What are you singing? This is going to copyright news, copyright news, copyright news, copyright news, copyright news. Anyway, copyright news is a small selection of news items. Yeah, that's good. Take it away. So first item. So, yeah, we mentioned that Greg was very integral in helping us run the Ice Pops conference, which we ran in July 20th of July. Wow. It was, yeah. It seems a long time ago, but it was a fantastic event. Yeah. The University of Glasgow. There is now all the presentations available. So if you weren't able to join us, you can have a look on the website. We've got our keynotes. We've got their slides. This is Chris and I with Amy Thomas, who was one of our keynotes, talking about copyright and video games. And we've also, we've got a couple of really quite amusing photos coming up there. We treated people that on both days didn't come about incredible performances. First day we did some country dancing. Yeah. Yeah. But he was backed by an amazing Salomon Orchestra from Glasgow University, including the director of Create the Research Centre, Martin Kretschberg, who I think was just fantastic. And our professional musicians. They were a great band and there was a copyright theme to all the music that they played. Yeah. And then there's, look, you and me dressed as Mario and Luigi. Yes. Which was another highlight of the conference and that's us on the closing panel with Professor Nick Whitton, who was our other keynote. He talked about Blade for Money. And yeah, we had a really great interactive session with her. But thank you to everybody who presented the ice pops. We've got lots of resources up from everybody who did lightning talks. We've got resources from people who are part of the World Cafe. And if there's anything that anybody wanted to share, any of the speakers at ice pops who haven't yet sent anything to us, just let me know. And I can continue to update that website. And to say thank you for everyone provided feedback as well. Yes. We've got some really great feedback. People still seem to like it, which is nice. Yeah. I think the next item, I think, Elizabeth, thank you very much in the chat I see that you mentioned what we're about to talk about next. Shall we go? The next slide. Oh, I've gone too far. There we go. So yes, the reason we were in Coventry is because we were at the old conference where, what do you want to explain what it was that we were there for? Yes. So we, well, we attended the 30th anniversary Gala dinner of the Association for Learning Technology. We did know that we had been highly commended for leadership in digital education. And we didn't win the award, but the judges felt that the work that we've been doing running these webinars and chairing and setting up the Copyright and Online Learning Special Interest Group was worthy of commendation. And so this is us with our certificates. I've actually got my certificate here and I can wave it when we're on camera properly. But yeah, we're really delighted there's further information on the old website about all the award winners. And I'm really proud of the fact that City University, my university did clear up quite a number of the awards. So the winner of the Digital Leadership Award was Deputy Director in my department, Julie Boats, and the Digital Education Team of the Year was also City University Digital Education. Yeah, but it was really great to be celebrated and we got to go and talk about Copyright again. We did. And to thank you to all of you that keep tuning in and that have contributed, of course, because we're the hosts, but it's you guys coming and talking about your areas of expertise and practice that makes this actually work. Yeah. So, yeah. Yeah. Thank you very much indeed. Yeah. No, thank you everybody. Yeah. There it is. Yeah. Yeah. So I spotted this actually just yesterday, I think this came out. So our friends at Create who are always producing these fantastic working papers have got one on Generative AI. So I thought that one was really relevant to the topic of today's webinar. Priorities for Generative AI regulation in the UK. I'll just pop the link in for the chat and let us say that one's literally hot off the press came out yesterday and it's going to be well worth a read. So, but yeah, I think that's it. They're all the news. I think so. So we are now delighted to introduce our guest speaker today, Alex Fenlan from University of Birmingham. Alex is the head of copyright and licensing there and has been working in the field within our community for many years. And as we know, there's been lots of conversations everywhere at the moment about artificial intelligence and particularly generative artificial intelligence. And Alex was kind enough to come on to the share with this copy seek his blog post that he wrote about considerations for researchers at Birmingham. And so we said, we jumped on it immediately. Alex, please come. Can we talk to you about it? And today, I think it's going to be a really good opportunity to sort some of the sort of red herring bits from the things that really things that really we need to know about in our practice as copyright advisors. So Alex, really great to have you here. Thank you so much for joining us. Double check we can hear you can't be I think we did check this before we came on but It should work. I haven't touched anything. I'm just going to get your slides up. Thank you very much. There we go. And so Alex over to you the floor is yours. Yeah. Thank you very much. And thank you very much Chris and Jane for inviting me to come and speak to you to everybody today. It's been a while since I've been at one of these. And it's good to be here good to be back and a good morning to everybody. So I'm going to crack straight on and get really into the to the nuts and bolts of the conversation that we want to go through this morning. I'll run through some slides but hopefully as Chris and Jane said there should be some opportunity to have some discussion at the end because I think really the the interesting stuff will come out from that conversation. I didn't really have a chance to look at the create report that was just mentioned just then but hopefully aligned with what we're going to cover in in the next 20 minutes or so. But yeah, we'd probably recommend reading that one to just update a knowledge and information to make sure that it's where we're moving in the same path. So we'll talk briefly about the context that we're operating in at the moment and activity that's taking place across the sector currently will pause and just spend a minute or two talking about the activity that we've been doing at Birmingham. And then we'll really get into the considerations that corporate advisors people who are the copyright gurus copyright expert and for institutions need to look at them to be aware of. And then heaven forbid we might even talk about some non copyright related issues which are equally important. I've added a couple of bullet points there at the bottom of the slide to say that biases is a really important thing to talk about and to be aware of, but we simply didn't have time to cover it in any great detail in these slides today. And neither will we talk about the carbon footprint and the impact of, you know, vast amounts of computer programming and the socioeconomic impact of AI tools being trained using labor from the global south that's a really important issue that's not covered in this conversation needs discussing as well I think. And then the naughty issue of hallucinations I think we might touch on that slightly but yeah something to be aware of that sometimes the references that these AI tools generate are not necessarily always accurate truthful at all. Okay. So, we'll start off by talking about the policy landscape, and we colleagues will be aware no doubt that the UK government has this ambition to be a global leader in AI and to make sure that we are a global superpower when it comes to AI knowledge infrastructure and creativity research and those sorts of things so the government has been very hot on AI investing significant sums into developing centers and resource and expertise, really trying to push the way in that particular field. We also have various sector bodies trying to to engage in the AI space as well, trying to support their respective communities making sure that they are aware to speed and utilizing the benefits of the AI can generate for their particular communities for our sector, you know, just a very active in this space RL UK and the Russell group with their recent release on their principles for education that we'll talk about. In terms of the legislation. It seems like we've been in perpetual consultation for the last three or four years with the IPO looking at legislative solutions for for for the AI conundrum. And we'll talk about that in a minute to but really on the ground what we're seeing probably since the start of 2023 probably post Easter, I think, we've really seen an increase in institutional awareness of generative AI tools, especially in relation to learning and teaching, and especially in relation to the use of those tools by students. I think there's a real concern within institutions that that students are going to be using some of these tools to generate their assignments and submit their work. We do know that lots of our academics are lecturing staff are engaged in trialing and testing and experimenting with some of these these new tools are available. You know, running their, their questions that their assignments through them and seeing what answers are generated, you know, asking them to generate lesson plans and assessment criteria and things like that. So, we know that there's lots of activity taking place certainly our place at Birmingham and I'm sure that's that's the same up and down the country. We have a new GAI channel that focuses on teaching and learning that was created in May, I think, and that's got over 200 around 200 members at the moment in the synthetic community, talking about all these different tools and models and methods that they're experimenting with. And we can see, I mentioned a minute ago the Russell group guidance. That's really been the trigger I think of the last couple of months at the early part of the summer for institutions to update their, their plagiarism policies, their guidance for students and for lecturers to try and deal with this threat or perceived threat at least of generative AI and the impact that it's going to have on learning and teaching. So, the National AI Center, JISC's National AI Center updated one of their policy guidance a couple of days ago, and really to say that this focus on GAI is leading to a re-evaluation of assessment practices and that's probably a good thing I think possibly long overdue. And really coming back to the copyright aspect of it, the legal aspect of it, I think it's fair to say that progress is slow, certainly in the UK. We've seen a shift from the UK IPO moving away, UK government moving away from a broad TDM exception that was announced 18 months ago, 12 months ago. That's been completely rolled back from and we're now looking at the codes of practice, voluntary codes of practice, codes of conduct rather than a legislative solution in the UK at least. The EU have their AI act, which is currently going through trial over the moment, it's going to be interesting to see how that turns out when the consultations and the discussions finally reach their conclusions, but that might take some time away. We also know that there are numerous copyright and privacy related lawsuits that are ongoing in the UK, US and EU, and it's going to be really important for us to keep an eye on those to see what the outcomes of those are, because they will have some real impacts and maybe to some of those questions that will come on to in a minute. I think it's fair to say, noting the irony, of course, that there is a lot of noise and a lot of hype around the benefits and the problems of AI at the moment, there are lots of lots of online seminars, webinars, there are lots of articles, talk pieces, comments and blogs talking about it. It's a very hot topic, as we said at the moment. And it's really hard to stay on top of that news piece, to stay on top of what we should be listening to and trying to get our information sources from. So I think, you know, events like this, forums like List Copy SIG are really important for us to be able to share and post information, articles, reports, things like that. So we as a community can try and stay up to speed with what's going on and how it impacts on our sector and the advice that we provide. So just going to pause for a second and talk about what we've been doing at Birmingham, really, really briefly. So, you know, we've been aware of the TDM exception in UK law since it was introduced in 2014, and we've really been working with our academic community to try and make sure that they're able to leverage and utilize the benefits from that exception within their research activity. So we've been supporting them with various queries over the years, building webpages, attending events and talk to them about it for a good long time. And really, that activity has very much been focused on them using, you know, library data, third parties, alternatively source data within their research activity, within their non-commercial research activity. We haven't really been involved in helping them or even thinking about building datasets that might be publicly available, building tools that might be publicly available, in any way that is similar to the types of tools that people are talking about and using these days. It's really been non-commercial research focused rather than anything that we've dealt with over the last 6, 12 months or so. We've been engaged in a couple of projects ourselves, so we're time to do an archive data mining project started in 2018, looking at how we support our researchers and provide access to datasets that we have procured for TDM purposes. And then we had a digitized to mine project that we've been running this year, which looked at how our digitization service can digitize materials specifically for the purposes of TDM, AI, NLP, you know, big data type research methods and what changes and processes are needed to be able to support that. Really, all of this journey has been about us trying to equip ourselves with the skills and expertise, knowledge to be able to critically analyze and assess AI tools, and really to build a trying to try to do, let me put my teeth in, to try to build a joined up service to support digital research, digital scholarship, and that includes AI, TDM, NLP, those types of things. In recent months, we've been working with our Higher Education Future Institute, HEFE, as they've been the ones leading on the guidance for GAI in the teaching and learning space. We've been working with our academic skills team and with our research skills team to try and update and tailor the support that they provide to their particular cohort, and that's been really interesting for the last few months. And then I've got a gritty or two shot of our Dubai campus that's going to give me time to have a quick sip, and then we'll move on to talk about the issues that we're going to explore today. So, what we've seen over the last few months in working with our academic colleagues with our lecturing staff is very much the focus is on how these tools can be used, how can they be used to benefit their teaching practice, how can they be used to supplement their assessment techniques, how might the students use them to write and to generate assignments and the issues that are associated with that. Every time one of these issues comes up on the Teams channel, I jump in with a copyright and licensing concern, and one of my colleagues mentioned to me the other day that there is a very real risk that some of these key questions that we think are absolutely essential, that horse may have already bolted, but I'm going to run into the field and try and wrangle it and get it back into the stable and have those conversations again. I think luckily for us within this community, I think some of the issues that we're going to tackle are not necessarily that new or unique to AI. I think it's not a case of the emperor's new clothes. I think we do have a level of familiarity with some of the core issues that are concerned here, but the lens might change slightly, the focus might change slightly. I think there is still that need that the colleagues in this community and elsewhere have been pushing for those basic copyright related literacies. And I think that AI literacy, this new term that's being talked about, will become in the form of a core element of digital and information literacy. And I think for us as copyright advisors and the copyright community, we really need to look at those core issues that we've been dealing with in most of our queries. So looking at the permissions, the ownership, the authorship and how exceptions and licenses and limitations might impact on uses of material with AI tools generated by AI tools as well. So we'll spend a few minutes talking about the six topics on the slide and we'll go into the first one, which is training data. So this is the, the, the active hard thing or scraping data, various different sources that's then used to test and train the algorithms that sit behind the generative AI tools. And there's a real question that we face a moment around the sources of that data. The tools are commonly used, you know, in vogue at the moment. Some are started off being quite open with where they source their data from and then we've seen over the last few months that that has been shrouded in mystery. There's been some ambiguity. There's a certain level of a lack of transparency that's that's involved in those data sources. Some people say that they're they're kind of black box tools, hence the flight recorder on the screen, that we don't really know where their data sources are coming from. And because of that, we don't really know if the data sources that have been used to train these algorithms to train these tools is legal. If they come from legal access, if they're based on licenses, if they're based on exceptions, we don't really know the mechanisms by which the data has been harvested. And that lack of transparency causes a concern, because there's a risk then that if the data is has been harvested illegally, that the sources could be problematic and troublesome for us to use and support. I think what we'll see from the lawsuits that are ongoing across the various jurisdictions over the coming months, he says, hopefully years is really a deeper understanding of what these tools do with the data when they go and harvest and scrape it, whether there is a wholesale copying, you know, whether they retain that data, they repurpose it and resurface it within the outputs, or whether what they're doing is more transient and more incidental. Whether they go and read something, analyze, take the statistical analysis out of it and then then they're done with a particular data source. And I think that's going to be one of the key issues that will come out from the conversations and the lawsuits over the coming weeks months. So then we move on to input data. So once the modular has been trained and has been developed and released, the users are then able to go in and add input and add prompts into the services. And of course, before you get access to the tool, you have to sign up to access it, you have to agree to the terms of conditions, you effectively have to agree to a license. And again, that will be more than familiar to colleagues on the call. And I'm sure we will know that that people don't read them. And even if they do read them, the chances of them understanding the nuances and what their means and obligations are extremely thin. And that problem, because there are lots of terms of conditions within those agreements that we need to be aware of that we need to be cognizant of. I pulled out just two of the key issues for this particular audience on the slide, really. And that is to say that when a user signs up, they will invariably sign up to say that they own the inputs that they provide. So that any questions, a text they put in, any images that they put in, they sign to say that they own that material. And that's fine on an individual user personal basis. That's no problem. But as an employee, as an institution, there was question marks about who actually owns that material. If I know from my personal circumstances, I don't own the material that I produce as part of my employment at the university. So then can I really legally sign to say that? And that's a question that doesn't seem to be answered or addressed in guidance or the conversations that we're seeing at the moment. So following on from the ownership question, invariably the user will grant a license to the platform, to the tool for it to then use that content for its purposes, for the purposes of providing the responses, but also possibly for the for the purposes of training and developing the algorithm, the tool further. And again, that's a license that the user is granting to the tool to the platform. And if we as individuals own that material, then we don't have the authority to be able to grant that license. So then, who should be signing those terms conditions? Should it be individuals in their personal capacity, or should it be the institutions that assign them enabling that use to take place? This is a conversation that we should be really familiar with in other aspects of what we do as well. So that's nothing new for us. We know that some of these tools have opt out. So in theory, you can select to say that I don't want my training data to be retained. And it's going to be interesting to see how effective that actually is and whether, whether that is the case or not. It's a challenging one, but something I don't think we're unfamiliar with necessarily within this particular audience. I put exceptions as number three on the list. And I'm sure colleagues are absolutely familiar with the TDM exception in Section 29A, the Copyright Designs and Patterns Act. And I'm sure we will, we will be familiar with what it says about this being limited in scope to non-commercial research activities, using computational analysis on content that we have legal access to. So we will be aware that there is a no contract override in that clause as well that talks about if there are licensed terms that prevent us from doing TDM, we can use the exception to override them and engage in that sort of activity. We also know that most GA itals are not non-commercial in nature. And that immediately poses a problem that they could be infringing under UK law. If, depending on what they do, it is viewed to be to take a to take a copy and therefore to be an infringement and I know there is an alternative school of thought of that. Whether there is any copying, whether it's transient or incidental. As I mentioned earlier on, if we look at the Copyright in the Digital Single Market regulations in the EU, we know that Article 3 and Article 4 cover scientific research for by certain heritage and educational research institutes in Article 3, and then commercial uses by everybody else under Article 4. Article 4 contains an opt-out which says that if a rights holder does not want their content to be mined, then they opt-out of that particular provision accordingly. What we know, I mean the TDM exception in UK has been around since 2014 and we know that there are some problems with it. We know that technical protection measures still a barrier and we know that trying to get rights holders to give us access is challenging in certain situations. We know that there are problems with data sharing, the ability to provide a mined dataset for data mining purposes, a harvested dataset with collaborators either across institutional boundaries or indeed across international borders is particularly challenging. And of course we know that neither of the exceptions in the EU law refer to anything around sub-licensing, the ability to grant permission to third parties over and above what it says in the exception. So we know that that's a particular challenge. And then when it comes to users, we will be acutely aware of non-commercial research, private study exceptions as well as the illustration for instruction exception. Again, I think that's probably fine, using these tools is probably fine within those particular remits, apart from the downstream licensing part. Again, none of those exceptions to my mind grant the permission that the platforms seek when you upload third party content. If it's your content, that's fine. If it's not, then there's a potential risk to be aware of too. Let's move on to to outputs. And this is a really interesting question around whether or not the V app generated by these GAI tools, these AI tools, whether they are copyright works in the sense of the legislative perspective, whether they are original or creative enough, and how that creativity process that originality of thought, how that interacts with the technical ability, the technical capacity capability for the program to generate something. We deal with concepts like the spread of the sweat of the brow of the author, you know, the author's intellectual, independent intellectual creations, these are, you know, central tenants to what originality and copyright are. And that's potentially a problem when it comes to AI because there is a human author there, and we'll come on to what the law says in a second. If they are copyright works, then what are they made of? Are they pure whole reproductions of original content of original copyright works? Are they reproductions of substantial parts? If they are, then there's a chance that they're infringement. Are they derivative works? Are they adaptations? We know that certain sections of the sector are using these tools for translation. There was copyright experts that translation adaptation are right to reserve by the owners, the rights holders, and that's going to be a problem that we need to deal with. And then we look at what the situation is with the UK versus international legislation, and whether the use of by these tools could be viewed as transformative in the US, and we've got Richard Prince and his use of his arguments around transformative reuse. Noting that he's fairly recently lost one of his cases around the Instagram posts. So that's going to be interesting to see because we could see a whole suite and raft of arguments in the cases that talk about transformative reuse in the US, and then of course how that impacts it does on the UK perspective. So I mentioned we talk about UK law as well. And if we assume that the works generated by AI tools are copyright works, then in the UK we have the provisions in section nine that talk about the ownership of those works as well. So that the author will be the person by whom the arrangements necessary for the creation of the work are undertaken. And we know that the legislation grants a 50 year duration for those computer generated works as well and we know that they are, they are works that are where there is no human author. But we also know that the UK is somewhat out of step with the rest of the world. And we know that again the US in particular take a view that that because there is no human author, AI tools, AI outputs might not necessarily have an owner which then might place them in the public domain which is going to be an interesting challenge. So how that plays out over the next few months is going to be really interesting for us to keep an eye on. And then I was taking the dog for a walk. And I thought, it's been a long time since I've done a copyright specific training session. And every time I used to do one a few years ago pre COVID, I had to include the copyright monkey, and had to include the selfie monkey. And really, if there is no owner of the, if there is no copyright in those works, then I can see a situation where we're going to be talking about the copyright monkey again. We're talking about ownership and human authors and there are parallels there with AI authors. But I think if there is copyright in it and we start untangling what the legislation means about, you know, who made the necessary arrangements for the creation. The user, the person sat there putting the prompts in, and we know that there's this whole field emerging at the moment called prompt engineering, where the quality of the input impact on the quality of the output and how you use that and deploy the tool will significantly impact the quality of those outputs. So if you're putting significant effort and thought and sweat labour and judgment into the inputs to use the tool. Does that is that sufficient for you to get ownership ownership of the rights of the outputs that come out. It's going to be a really interesting question, or is it actually the AI tool itself, or is it the company that produced it the individual organizations that produce those tools. And it's really going to be a really interesting time for when those conversations when those arguments come to fruition. Yeah, and, you know, can monkeys own copyright can AI own copyright. It's a really interesting and naughty conversation and, and you know, it's a really nice picture so ways like to include that one. The last one of my six points I wanted to talk about in terms of a copyright specific queries for us to be aware of is this balance between an interplay between infringement and plagiarism. You know, if we talk about copyright infringement in the use of some of these words, for example without permission and without coverage under an exception. You can use other words you can rephrase and paraphrase, and you might not infringe the expression that's protected by a copyright for plagiarism. However, it's slightly more slightly different. If you copy the idea without the attribution, then there is a risk that you're going to be guilty of an academic plagiarism infringement, even if you use those different words and it's going to be potentially problematic. So this is where the idea and expression dichotomy comes in that colleagues will no doubt be familiar with. And to what extent are the ideas separated, separable from the text that's actually used within the expression itself. It's going to be really difficult to disentangle some of that stuff in certain situations. And especially, that's going to be the case where GAI tools do not reference where they're sourcing material from. And we know that lots of the most popular tools simply don't reference where they're sourcing their information from. And the lack of referencing is going to be really problematic for our students. How do they know who they need to concite and acknowledge when that information isn't attributed at all. And indeed talking about hallucinations briefly, references that simply do not exist. And I'm sure our front desks and inquiry colleagues going to be inundated by queries over the coming months when students come back. And I've got another gratuitous shot of campus just for me to take a drink and then we shift focus to some of those non copyright specific concerns. Okay, so one of the big ones I wanted to talk about first and foremost is the really naughty and tricky issue of privacy and personal data. We know that some of the big tools have been subject to various complaints about how they process personal data. You know, Italy have launched complaints and banned chat GPT recently, although it was restored. We know that Austria had issues with facial recognition and biometrics data. And we know that Ireland and the European Commission are taking increasing interest in some of the privacy policies and some of the way that some of these tools handle personal data. They're really looking at under what legal basis are these tools processing that personal data, how is it collected retained. And that's going to be a really interesting conversation. You know, if you look at it from a personal perspective, if you're uploading, you know, information, personal details on one site. You might not necessarily expect it to then be harvested and reused and repurposed under a completely different website format tool without your knowledge. And that's a real concern for us as individuals, but also within our professional capacity as well. And what I'm seeing on the ground, certainly within our institution is when you ask about data protection, when you ask whether your data protection impacts assessments are being completed, they're not necessarily following those things through. The same process that we might rely on in our ethical clearance might not necessarily be being transferred over to our teaching concerns as well. And that might be considered problematic. And then there's the issues around referencing and how do you reference the GAI tools within your delivery, within your assessments, within your assignments as well, which could be potentially challenging. Alex, I just come and let you know, we've got three minutes left in the session, so I think we want to leave time for discussion at the end just to kind of figure out where we are. You carry on. I'll rattle. Yeah. Thanks, Chris. I'll rattle through these if I if I can. Responsibility, reproducibility and consistency is an interesting problem. There's been certainly in the early days of these tools being launched, and it might be less of an issue these days, but results would vary by, you know, day by day almost hour by hour. So you can put in identical prompts and get completely different responses. And that's a real challenge in terms of the research space, because that that reproducibility being able to check and validate to test and repeat results is a real crucial part of the research process. And if we apply that to the teaching space, and that's going to be increasingly challenging, because ultimately it could impact on a student's results, it could impact on a student's grades, and that's going to be a real challenge. And there is a real risk that if these tools don't provide stable results, stable consistent results, then this reproducibility crisis that we're facing in the research space could well extend into the teaching space too. And that's something for us to be aware of them to raise concerns with. Sustainability and equity is also another one. Lots of these tools are currently open, but we are seeing, you know, freemium models, full premium subscription models coming on board where you can get access to the latest version, latest tools for, you know, the princely summer of $20 a month or whatever it is, whereas the free version offers an outdated version with limited capacity capability. So those that can afford to pay the subscriptions, then they will get a better experience, better insight than those that cannot and that's going to be a problem. And in a world where there are no institutional models, institutional subscriptions at the moment, requiring our students to pay and access these tools could cause problems with the CMA, could cause problems with the OFS as well. And also, you know, these tools are subject to a huge investment from Silicon Valley. Silicon Valley moves very quickly what might be invoked today might be old hat tomorrow. And there's a chance that the tools that we are using today could disappear from public view completely. They could be hidden and incorporated into other products as we're seeing with some of these tools already, or they could simply be dumped and no longer supported in any way, shape or form. So that's a real issue for us to be aware of. I'm going to whistle through these next two. So pre publication submissions, if you run your, your article, your grant submission through one of these tools, and, you know, for a grammar check or for for appraise a concise version of it, something like that. Then, then because of the license that you grant to them there's a risk that that knowledge that information that you communicate could then be resurfaced to others without attribution and could prevent you from securing that grant or securing that publication agreement. And that could be a real problem. And if we look at research within a research context, we know that if you're trying to pattern something non confidential disclosures is a real problem, putting information into the public domain is a real problem for for patenting and lead to a pattern being rejected on the ground is already within the state of the art within the common knowledge. And similarly, if we look at commercialization of research is a big push for research funders to translate research from from the bench to the bedside and back again, as they say, lots of research in non commercial research in UK into AI. And the question mark about when, when do these tools that might be being developed, when do they shift from from being non commercial research into commercialization translation activity, and what happens with all of that training data and all that harvesting work that's been going on in the back end. So a summary. I think it's really important to read up on some of those issues around those biases, the social and economic climate factors that can be key and essential to understanding how GAI tools, and they remain to be explored and I would would advise colleagues to go and explore some of the conversation that on those. There is lots of noise around AI there is lots of hype at the moment. But I think our core knowledge that we have and that we exploit on a daily basis is really essential. It's going to be really important to keep an eye on the code of practice, the IPO generating the AI Act is going through the EU at the moment that's going to be really important to keep an eye on, as is those cases that are going through and breathe. I think that's me. Well done, Alex. Alex, thank you so much. Yeah, we're going to share your slides off to thank you so much. And we were just having a bit of a chat here and saying, there's a lot we could be talking about. And we're wondering whether and we're wondering whether next month we do a more sort of discursive session as well about copyright and AI and continue this. I think so. We've got some time now. I think we can start with some questions off and perhaps dedicate next month's whole hour discussions about this. I think that's it. I think that is very much a whistle stop overview of some of the issues and there's lots of nuance in each of those six issues that I mentioned. And there's probably, you know, 15, 20 issues that we could talk about as well. So having the time and space to be able to explore those would be really, really interesting and really, really valuable, I think. We'll advertise that in early October, I think. I mean, I've done what would be useful and we do have a bit of time. I mean, that was a fantastic, fantastic presentation, Alex. Thank you so much. I mean, the areas that I know we didn't have some questions. We have questions about Naruto and the monkey selfie so about ownership digging in to that a bit more. And therefore questions about citation about what we say about material is produced by AI. I've got some sort of broad thoughts about how as a copyright advisor. Work through all those issues remaining in touch with all of our colleagues who are thinking all those broad issues and how do we make sure that we input our expertise in a way that's helpful. Oh, yes, absolutely. Rather than coming in and flooding that conversation with all the layers of complexity that we know exist around ownership of potential property and the terms of conditions. Generating some kind of coordinated institutional response. I think that's really interesting stuff about what we see in pushback against ideas that traditionally those of us in libraries, cultural and educational institutions have been pushing for for many an openness, flexibility in the law, flexibility for text and data mining. I don't know whether we're a crucial moment for sort of working through those and not having a sort of a knee-jerk reaction, but also thinking quite carefully. Your metaphor about whether the horse is bolted. There's a number of things that I think particularly around the terms of use that we sign up to. We all know that we sign up to terms of use all the time as a matter without thinking institutionally whether that's it. A whole load of things. Can we go back briefly then? I just do a whole load of things that we will return to you. But the question that was asked is about the monkey cell case. Evelyn has asked, did they decide that copyright would belong to the monkey, but monkeys can't hold copyright. That complex case wasn't it because there was another lawsuit that came in from Peter, the people against the ethical treatment of animals, which confused things quite a bit. Do you want to give a little bit of background to that? I think it's a really interesting analogy to draw on, because those conversations are going to be really relevant for when we talk about the AI activity and how that's generated. And I think that uncertainty is going to raise again through the coming months and hopefully not years, but it will be years. And yeah, it's about who created the work. Is it the machine itself? Is it the person that generates the prompts that then lead to the output that's generated by the tool itself? Or is this algorithmic beast going to generate a legal personality and therefore be capable of owning property? It's a really interesting and knotty conversation that we need to keep an eye on because it could be potentially hugely impactful. And no, I didn't answer your question. No, no, but I mean it gets to the heart really of what kind of philosophical questions now people start to ask themselves. And we see this from the creative industry as well. Don't be about what needs to be human and what creativity, true creativity actually is. And actually copyright really is at the heart of that. I think it's not that surprising that in quite a number of stories that are coming out about AI copyright is being flagged up or something. I think there's an element of it where that could be like scaremongering about content owners who were concerned about their content being ingested and used in ways that they don't want. But I think there is some genuine concern as well about, you know, if I create a work, ask an AI to create a work in the style of a living author or a living artist. You know, I'm kind of... Well, I guess the question there is what is the difference between asking an AI tool to create a parody or pastiche or something and a human being doing it based on their own experience of the work? Is that question you've been asked at it? No, not yet. But I think hopefully I won't get asked that question, but I'm sure I will. It's, yeah, absolutely. It's that same issue, right? And it's being able to identify what is protectable via copyright and what isn't. You know, I remember being in conversations with the RPO some time ago. And, you know, they started bringing in notions of passing off from trademark and trademark law and stuff like that. And that's perfectly valid, I think, but it's concerning, like, because copyright law is not trademark law. It's not common law in that respect in the sense that passing off isn't so it's different. And that's potentially problematic as well when it comes to copyright infringement. You know, we all know the the idea expression dichotomy and that fine line sometimes can be problematic. You know, what is a copyright work versus what is the idea? Yes, yeah. Yeah, I see Andrew, let's put a comment in the chat there. And it's about ease and money in that scenario. Consider as well. It's getting an AI to do it than another person. So I think we should return to when we come out of the discussion. We talk a lot about how to avoid claims of copyright infringement by not reusing, say, photographs that you found elsewhere and not have permission for. Should we, as copyright advisors, be saying, oh, well, just get an AI to do something. It'll come up with something that's completely a different photo. It's original, but it's ultimately the same. And I think that's it's not an easy question to answer. I don't think I would feel comfortable by saying, hey, just get an AI to do it because of all these issues you've raised. And similarly, we need to take it away. I think that's right. And I think black box nature of some of these tools, right? We don't know where they're sourcing the content from. We don't know the terms, conditions under which they operate. We don't know whether the harvesting and use of their material is legal in the first place. Right. So while I, you know, I understand the thought process that says, yes, it might be easier. You might be quicker to do that. There are those legal ethical questions that are greater than when you go in and you source an image from a particular website, you cite it properly, you use it with full justification under the exceptions. And that resolves, reduces the risk, I think, because removes the uncertainty. You know who the creator is. You can cite them accordingly. Yeah. That's a really good point. And I think so we'll return. That's kind of returning to the principles of how we've always managed to be a copyright issue. Absolutely. Which is why I start, yeah, which is why I start that conversation with this piece. We're saying, actually, those literacies, some of those basic skills we've already got. We use them on a daily basis. And it's just about representing some of that stuff within a particular context, which is bread and butter for this community. Absolutely. Alex, will you be able to join us? And we'll pick up some more discussions in the next month's session. We'll be over a suitable Friday early in the month. That's fantastic. Thank you. Thank you. And thank you so much for everything you've shared. I just want to go back to our slides because we just had a couple. Well, we just have one thing that we wanted to flag up. I think with a resource I've been working on at the moment. Sorry, I have put the link because you were going to put the link to your and Stephen's video. Yeah, so I'm on the Generative AI Working Group that we have at City University. And we are putting together a whole series of resources aimed actually primarily at students, although I'm doing a piece of work about staff training, about AI as well. And I've run a couple of workshops. I think Deborah mentioned huge amounts of the concerns are in her institution, as around plagiarism and academic misconduct, absolutely similar in many institutions as well. But Stephen and I were asked to make a video aimed at students about copyright and AI and to sort of touch on some of the wider ethical issues. It's not yet available for students, but we've got it on our media space platform so that anyone can have a look. I'm really welcome people having a look at it. And, you know, if you've got any thoughts or feedback, let Stephen and I know and say it's a suite of resources that arrange primarily students, but we do expect, you know, that staff will use it. Hopefully, I'm going to talk to the group about the licensing of these resources. So, yeah, that's just something I wanted to mention. But we talk exactly about some of those issues around, you know, the kind of overlaps between copyright and then when you stray into academic misconduct, and whether it does or doesn't help if you get an AI to create images, et cetera. We now, we're at time. We are. Future webinars. We have agreed just now that we're going to return in October to this very topic and delve into more detail. But we are definitely have another guest on a different topic, you know, member David Beals. Yeah. He asked about open textbooks and David spoke at Ice Pops and we know it's going to be a fantastic conversation that he did a really great vital talk. Yeah. So, thank you so much, Alex, for, you know, for coming along and all the work put in. Yeah. That was a really excellent presentation and I'm really looking forward to taking you to the conversation. Yeah, we've got such good feedback as well, so thank you everybody for being here today. I think we've just got... One last thing. Our one last thing. We're not going to play the David, yeah. You can explain this. It's nothing to do with copyright and AI at all. This is inspired by the recent interest in the Barbie movie. I hosted an exhibition in the town where I live in Kent on my... Look at this. My favourite... This is incredible. A Cindy collection in addition to many of the things I've interested in. I love my Cindy's. It's got... It just came down last night but I just thought I'd share some pictures. So we have some close-up. Yeah, come on. Can we have a close-up of Cindy in action? So what's going on here? Do you have her horse? She's having a little chop around the paddock with her horse. She's keeping fit while getting all her chores done in the middle one. Over in the deck? Yeah. In the garden? Yeah, that's why I started on Cindy. She's been out for a picnic on a scooter. I don't quite know how she got the pram and the baby there, but Cindy is a Wonder Woman who can do everything. Yeah, it's been really popular on social media as well and on local. And some Cindy, Facebook, national groups as well, or international groups. Yeah, Cindy in the kitchen, in the bathroom, and she's getting ready to go to the ball in her dress designed by the Emanuels who designed this van as their dress. So I was very excited. I still had all those things in my loft. So there we go. Something you can take away with you. Yeah. If anyone wants to see it closer up or, you know, drop me a line. I'll happily send you. I've got an article written about it. Right. Thanks so much for coming everyone. We will see you next time. Yeah. Thank you very much.