USERrights 2026

And then came AI. … Copyright is again on the front pages. … Fears of the displacement of creators and workers and increased demand on open content providers and reports literally of the death of humanity are shaping our positions once…

And then the most important one, I think, is the money, is following the money. So that's the shape, I think, of things, is that we're the people at the bottom supplying the inputs and we're the people at the top requiring the outputs,…

India is, of course, having the largest number of voters in this world. … So someone came up with this wonderful idea that maybe we should use AI for cleaning up the voters list, with a net result that millions were thrown out of the…

I see sovereignty in terms of resilience, not autarchy. … what we are talking about is open source, and is by definition and by design multilateral, and is not about closing, but it's about opening.

…when local AI developers approach these entities to have access to that, they will assert copyright, but it's not really about copyright alone. It's more about the data extraction fear. Once we let you access this, what will happen to…

But right now with Trump, it's clear that Latin America is the backyard of the U.S. And therefore, we are pressed with its policies, among them innovation. So innovation is the rule for discussing AI in Latin America.

There is also an emerging issue of residential proxies, which is basically scraping that happens through private people's devices, and it's undetectable as massive scraping but has the same effect, right? … And there were cases, not…

…what kind of an exceptions environment do we think is appropriate in this current moment? If we were going to take one step forward, would it be the same step that we would have taken 10 years ago…

So in a nutshell, the fear is that AI systems only speak US American and Chinese and don't speak all languages that we have in the EU.

Every negotiation starts from scratch, and publishers expressly tell us that when the University of California is not at the table, other institutions simply accept whatever terms are offered. … They can't push back, because they lack…

…the uses that we can do of AI are also socially positive and desirable. … there are also situations where they can be extremely advantageous for the advancement of, of knowledge, and therefore the question becomes much more a…

So Slovenian implementation is a strong case for public interest AI. … We are aware that our small language, we are two million living in Slovenia, hopefully more speak this language, that our language will not survive if AI will not…

…but the right to do these things is the thing that should carry us through and help us to beat all of these obstacles that are put in the way, … And if we don't frame it as that, as a positive, enforceable, actionable right to do the…

The idea is to get everyone to keep creation flowing, to keep people interested in creating. So if you know that you can be an artist and you can be a musician and you can create and you can still pay your bills, this is a huge…

And looking from that perspective means that maybe creativity should be protected and remunerated not because it would compensate for some harm, but because we need to sustain culture, knowledge, as a common good essential for society…

So I guess what I'm trying to argue is that if we invoke human rights to justify AI remuneration, we really have to be consistent and follow that argument through, and ask who actually gets paid and whether the creators really benefit.

We have beautiful solutions on paper that might not work in most countries. … It's not sustaining creativity in Europe. If that was the case, I probably wouldn't be here. So we have to think of models that work in countries with weaker…

And I think that Professor Dusollier's point about copyright as part of a cultural sphere … and landscape really resonated with me, because I think if we use that model, … we can think about how we regulate that sphere in a way that…

This is not a selective process; people have not taken, "I'll take this and I'll take that and I don't take that." They have taken everything that was up for grabs, and that is what made this technology different. … So trying to solve…

We all agree, maybe, that copyright or intellectual property is a private right. … is there a responsibility for governments to regulate those private rights?

So the broader question is: is there, within more long-standing principles of how states behave to one another, an argument that you must not allow your territory to be used for harming others? And I think that's a fairly solid…

We get that in Canada too, largely from Canadian companies that want government handouts to create data centres and argue that what they're going to do is create AI sovereignty. Respectfully, most of this is nonsense. … There is no…

And these burdens are definitely not evenly distributed. … Because large firms, the large tech companies, can absorb the cost. They can reorganise their infrastructure. They can restructure their contracts to align with the ecosystem.…

But it may also be the case that maybe there's the risk that the creative industries have cried wolf one time too often: like, Betamax was going to destroy them, and then the internet was going to destroy them, and then Netflix was…

We can't start a conversation of just AI and forget about why do we not have just digitalization. … Is it for everyone really AI access the question? Is it really what everyone needs? … Does everyone need frontier AI models? No.

The right to science and the right to research are two sides of the same coin.

I think there’s a vast and fundamental difference between the two that we have yet to explore thoroughly…

The author is one who causes to grow, the creator, the originator, not one who destroys.

Some US legal scholars have argued that a market-based, or at least a non-government-controlled, system of publishing, open publishing, open science, open access, in part facilitated by copyright or open license models, which depend on…

In fact, I would argue that the challenges posed by AI are amplified and made more vivid when one speaks about TK. TK is therefore an edge case. It's a hard case on which there's much to be said that is relevant to designing a…

So I see it less as a copyright problem and more of a larger-scale problem of exploitation, extraction, taking of knowledge and power. … But looking at more individual-level remuneration per work, I think, is an insufficient approach…

The main thing, and I'll be very blunt about it, is that oftentimes open access conditions that come with big funders basically violate previously agreed consent negotiations. … You see also big funders that come and say, oh, we must…

…if copyright governs a researcher's corpora of data for a research project, the copyright is biasing the research, yes. The corpora, the data that is used for research, has to be governed by the research question, not the copyright…

And the harms are significant. … The market replacement of the authentic art with this unauthenticated reproduction devalues the community-made artwork. But beyond that, the inaccurate representation, the misinformation, as well as the…

By voice I mean how do you have Indigenous peoples define and describe their own knowledge rather than simply have another speak for them, so having that authenticity. Agency is the ability to participate meaningful in decisions about…

The problem that we are facing is not so much that we lack ideas, we have lots of them, but that we need better coordination, and that we are currently working in too many silos.

…was that AI will certainly not manifest itself in the same way in the Global South as it does elsewhere, and that this will require different approaches.

Digital sovereignty; data governance and infrastructure; public AI and global market concentration; fundamental rights; copyright, text and data mining and AI training; and then lastly, and overlapping with what we just heard about,…

We had already kind of pre-prepared, Allan and myself had prepared a couple of research pillars, and we found them confirmed during that session by the presentations. So going forward, we will do some research in the following research…

So then at a certain point, all that matters is who once upon a time in the 20th century owned everything, and now they just get a cut of everything that everyone makes in the 21st century. That to me is a kind of disaster.

Ultimately, I am a human creativity optimist. I just think that whatever AI, and to some of the points, whatever AI does is never going to be able to replace sort of the human creative spirit. That sounds a little corny, but that's…

So I'm saying that's what we should do, or might at least consider doing, in this context: saying, yeah, you used AI, that was a choice; yes, you're still an author, but we have decided our copyright system is not here to incentivize…

So this is, again, something about users' rights, you know, how to make users' rights not just a means of defence, but, you know, a positive action or claim from users, saying, "I have the right to do that, so please remove your…

At this point, maybe copyright is obsolete for a world of AI, but it doesn't mean we don't need a form of legal protection that allows authors to remain authors, if expression is what makes us human authors.

Putting all these things together, we could really see a lot of public AIs evolving on a national basis or on a regional basis.

Because libraries around the world, the national libraries, not the research libraries or the public libraries, these are typically the national libraries, are getting these deposits.

And so it occurred to me that, and one of the beauties of sort of focusing on national libraries and their role, their potential role in developing public AI, is this is, again, as I said, outside of the copyright space.

And just to explain what that is: in most countries, or in many countries, there is a requirement on typically publishers and other creators of copyrighted works to deposit their publications with the national library.

The topic of AI makes certain countries and certain stakeholders very nervous about research limitations and exceptions.

There is a proposal moving through the SCCR to reunite discussions on the limitations and exceptions agenda…

We might want to circulate a letter or something and tell them we think it’s a good idea to keep research in the ballgame.

I mean, one is that at the last meeting, the research institutions part fell out of the agenda. So they're now focusing on the libraries, or the preservation interests of libraries, archives, and museums, and the rights of people with…

AI is predictive technology, and the more good-quality input it gets, the better the output would be, right? And so, we've given collectively our creativity, but now we've been asked to pay subscription, maybe even pay by sort of…

Now everything that is, you know, on the internet, on AI, what they are sort of feeding on is data. … And that data does include scholarly communication. So because of this, the politics of access to knowledge is the politics of the AI…

…how can we actively shape those conditions under which collections are actually accessed and reused by bulk AI systems without compromising this openness? It's an impossible quandary and paradox that we're trying to deal with, but…

Like, if you look at Reddit, if you look at all of the different social media, like, you know, Tumblr, LiveJournal, all of these things that have had huge communities. Like, I don't think people contributed to those because they were…

I know our legal counsel is very frustrated, because someone will rely on, on some terms of service and then all of a sudden they've changed, and now they've violated the rules. … how is anyone supposed to rely on what they accessed…

I really liked when it was framed as the defence of the commons, because we are in an environment in here in Geneva, we don't have an international knowledge organization that even looks at different ways: how to construct knowledge,…

If you're dealing with entertainment and big tech is the user, or if indeed it's sacred traditional knowledge... … And what you're doing is directly reproducing or mimicking, you might lean towards protection. … When you try and get in…

I think we also moved from the framework of defending to the framework of protection. And I think these are indeed a bit of different approaches, not only semantics, because protection, I think, is also about nurturing, right? This is…

…the national library in Norway is using the legal deposit to create a public AI. … we could really see … a lot of public AIs evolving … on a national basis or on a regional basis.

I don't think AI is merely an enabling tool like a typewriter or a camera. … it's a tool that can flatter, persuade, deceive, power seek, and even ultimately possibly control us. So how do we recenter the human in the knowledge…

And I do believe that sovereignty is a framing that this community owns. … what it really means once you unpack the marketing wrapper, is about the right of individual countries to make sovereign choices from a legal perspective for…

…at the last meeting, the research institutions part fell out of the agenda. … There's disputes within different countries about whether they want the globalization of research exceptions. … The topic of AI makes certain countries and…

So we want to continue to talk about openness, but maybe we need to redefine openness, and need to think if each of the elements that we have considered essential for openness can still coexist with large-scale reuse of the commons.

…all the very policy issues that we have been discussing the last days in relation to AI generally, consent, access, control, inclusivity, attribution, compensation, humanity, sovereignty, authenticity, they are amplified when we talk…

Aren’t we also users of AI tools? Aren’t we users of libraries? Aren’t we users of open license content and knowledge systems? Aren’t all users also citizens with an interest in self-government as a matter of right, not just interest?

…has to bring about just AI, because unless you’ve got a different form of AI, you’ll never fulfill that user right

We all seem to be in a post-copyright world... But when we look at where the frontiers of user rights are actually being fought, copyright remains our vehicle to do things.

It's like, we have fixed the holders of rights are there, then they use the rights, as is just being said, defend, defend. But maybe, especially in the field of AI, there's another way, not to look at everything as being IP or copyright.

Right now, we’re in this phase where we’re using AI tools and doing lots of fun and exciting things…

So let’s have that balance not only in terms of AI companies, but also in terms of cultural companies.

I don’t think it’s fair to say that this community is not so much concerned about users’ rights anymore.