Guardians of Scholarship: Librarians and Copyright in the Age of AI
Protecting and promoting copyright in the rapidly evolving context of artificial intelligence is a relatively new concern for academic publishing. But Christina Daouti is confident that librarians are rising to this challenge and adding new strings to their bows.
Did you know? Taking Libraries into the Future is also a quarterly webinar series for academic librarians. Learn more here.
Academic librarians are not only law-abiding but also extremely well-informed. Researchers have long relied on their guidance in order to navigate the choppy waters of open access and copyright. Meanwhile, the recent addition of artificial intelligence to the scholarly publishing landscape has made librarians’ contributions even more valuable. Publishers are also deeply indebted to librarians for upholding copyright standards and protecting them and their authors from the misuse of research.
Most senior librarians devote a great deal of time and thought to understanding, explaining and upholding copyright; but few have made it the focus of their careers to the extent of Dr Christina Daouti, who is the Copyright Support Officer at University College London. We asked Christina to explain her role and her views on copyright and AI. For this latest instalment of Taking Libraries into the Future, she was in conversation with Gold Leaf‘s Linda Bennett.
Linda Bennett: Please tell us a bit about your career and how it developed. How did you become involved in such a specialized area of scholarship? Is it something that has always fascinated you, or was it a case of picking up an opportunity and running with it?
Christina Daouti: I did not qualify as a librarian. I was a researcher and was awarded a PhD in Psychology. Then, 20 years ago, I saw an advert published by the University of Surrey Library, asking for applications for someone to take up a role to explore and support open access, which was relatively new at the time. It seemed exciting and I wanted my career to go in a different direction, but I had no idea that I would stay in this field.
Without wanting to pigeonhole open access and scholarly communications, I quickly understood that they share copyright as a core issue. Copyright is what connects academics to publishers, and to publishing activities both within the institution and outside it.
It is not common to be exclusively devoted to this area – and at first, I wasn’t. I became increasingly interested in copyright literacy and how it can be a key to opening up research. Over the years, as open research evolved, other areas such as rights retention and open licensing brought copyright considerations even more to the forefront.
My role in open research was very broad and copyright was only part of it. When, three years ago, UCL advertised for a copyright support officer, I saw the opportunity to focus entirely on this field. I really wanted the job, but I didn’t anticipate how interesting and varied it would be, especially with the explosion of generative AI soon after my appointment.
My job is also very diverse, even though it is all about copyright. It involves advising on publishing agreements, licensing publications and data, reviewing archive policies and managing copyright in dissertations, educational resources and websites. I also run sessions for various UCL audiences: developing playful and accessible copyright resources and games and, more recently, setting up a community to develop copyright literacy across UCL. It’s a very creative, innovative environment and this is reflected in the variety of projects I get involved in.
“In my view, AI might represent one step further than open science – but it’s difficult to assess where that step will take us.”
LB: Publishers really value librarians as excellent custodians of copyright. It would appear that academics often don’t have the time – or perhaps the inclination – to explore the detail for themselves. How do you explain copyright to academics, including their own rights as authors? What sorts of questions do they ask you?
CD: I spend a lot of time talking to academics – and you’re right, they’re not always enthusiasts. Many years ago, an academic asked me to make copyright “go away.” Researchers and lecturers are busy and this needs to be acknowledged and built into the training and support I develop. Having said that, copyright is becoming more directly relevant to academic practice and people are increasingly likely to understand its relevance before issues arise.
There are different audiences within the scholarly landscape which need appealing to in different ways. Most academics come to me with a problem, whether it’s about ownership, licensing or using images for theses or articles. But the real challenge is to convey the message that copyright is relevant to those who have not come across a problem yet. Copyright literacy is an invaluable skill to develop before issues or questions emerge.
The best approach is to anticipate the issues and offer education and support that is meaningful to diverse audiences – academics, students, researchers, professional services staff – while also acknowledging that there are cultural differences: different people have very different ideas about IP. I have discovered that how a session or resource is labelled and promoted is crucial. For example, a generic session on copyright in itself may not attract interest, but hands-on sessions such as “copyright for your research data” or “copyright and generative AI” will.
LB: Do you think that copyright for academics and publishers is threatened by artificial intelligence? Or are there ways of safeguarding it? And again, how do librarians – and particularly you, as an expert – address this?
CD: There are certainly ways of addressing AI, but the question is not an easy one. People feel in the dark – and therefore, threatened by – different aspects of AI, including biases, inaccuracy and misperceptions, as well as ethical and environmental issues. There is a lot of uncertainty about the effects of AI and these are not confined to intellectual property.
There are stakeholders in the AI debate, but it has been very much led by the creative industries, which have different aims, practices, models and even values from academic research. There are overlaps, of course, but the various sectors need to be addressed differently.
In my experience, academics feel uncertain about what happens if their research is used to train AI models. There is a lack of transparency which is not helpful here. Then it is a question of attribution. To what extent can publishers assure authors that their research will be properly attributed by AI? While AI models are evolving to link material back to the original author, the capability of different tools to address this is currently questioned.
Overall, most originators of content say they want some control. It therefore becomes a very emotional subject. Currently ways are being developed of creating more transparency and awareness. But perfecting this will take time. It’s a real challenge: I need to convey the potential benefits and opportunities of AI while also addressing the concerns. It is a fine balance to translate into guidance.
LB: What is your personal view on AI? Do you think that on the whole it is beneficial?
CD: My view is that AI can be a force for good, but transparency is key. It has tremendous potential to help disseminate research more efficiently and quickly, and also – despite the well-publicized instances of abuse – make it less biased. AI might represent one step further than open science – but it’s difficult to assess where that step will take us.
“Change takes place on many levels – for example, we need a more responsible apparatus for the assessment of research and better methods of dissemination, and publishers can certainly carve out a role for themselves in this new world.”
Like open science, AI can open up wonderful opportunities, save lives, offer solutions to complex problems and lead to unimaginable discoveries. At the same time, there is also the potential to cause harm. Much of the concern is about the speed and scale in which information is being processed. This makes us consider the differences between application by humans and application by machines. The regulations and checks that humans can enforce are crucial.
LB: How do you instruct academics about AI?
CD: I have recently developed a session entirely on AI at UCL, to encourage discussion around the issues mentioned here. I started by incorporating it into existing sessions on copyright. One department wanted me to talk more about the implications for their research, which meant more presentations and bigger discussions. Text and Data Mining (TDM) is also closely related to these discussions and the guidance I have created addresses that. It is important for me to understand and convey the nuances while addressing the contentious nature of some of these discussions.
LB: How do you see copyright developing in the future? And do you see the librarian’s role developing to meet the changes that will take place?
CD: We should advocate for and embrace significant changes to copyright legislation. Some of this is more than 300 years old and no longer fit for purpose. I think that UK copyright law, for example, should be reformed to recognize researchers as a specific user category that qualifies them for exemptions and legitimate secondary use.
This will take time, but in the meantime, copyright literacy – among information professionals, researchers, academics and students – is important. It’s quite a big development (led by Jane Secker and Chris Morrison in the UK and supported by so many others) and librarians are at the forefront of picking up on the interest in this.
Copyright literacy should focus on the flexibility copyright offers through exceptions and interpretation. We must highlight that copyright is important as a compliance issue. But it’s also important to understand the views of different parties – as certain legal cases currently passing through the courts demonstrate.
How we think about copyright is constantly being revised and needs to be continually adapted to the context of AI. The librarian’s role is to keep up with developments, digest information and work with others to address implications for research and education.
As a profession, we already master so many different areas – curation, education, advocacy, enforcement – and new developments in AI will further enhance the role of librarians. That’s why I’m certain we can succeed. Over the years, librarians’ roles have evolved to support systematic reviews, TDM, and now AI as well.
“As a profession, we already master so many different areas and new developments in AI will further enhance the role of librarians. That’s why I’m certain we can succeed.”
LB: Do you think that in the long-run authors will be better off for having engaged with AI?
CD: That depends on what “better off” means – it takes us back to the purpose of research. Is it to benefit the world as a whole? If so, there are immense possibilities with AI, despite the concerns. If “better off” is about the author’s reputation, recognition and remuneration, then regulations and appropriate technology must be in place. Royalties, patents and attribution all need to be addressed as well.
LB: How do you see the future for publishers taking shape within this context?
CD: It’s difficult to say. I think that the rise of university presses will continue to gather strength and we will see innovative business models emerge, such as Diamond OA (the model adopted by UCL Press). Discussions on rights retention and open access may make some traditional publishers feel threatened, but they have at their disposal many other ways of adding value. Change takes place on many levels. For example, we need a more responsible apparatus for the assessment of research and better methods of dissemination, and publishers can certainly carve out a role for themselves in this new world.
LB: Between all of these exciting projects, do you find time to relax?
CD: Certainly, but my tendency to think about ideas and concepts follows me outside work. I love reading and watching films but this, interesting and exciting as it can be, is not always relaxing! Music, cooking and just spending time with family and friends, preferably not talking about copyright and AI! This is what makes me truly switch off.
LB: Dr Christina Daouti, thank you very much indeed for this fascinating insight into the librarian’s role in copyright.
You might also be interested in this blog series
[Title Image by matdesign24/iStock/Getty Images Plus]