AI disputes are here. Are you ready?

Let's dive even deeper into that question of where do the interests and the demarcation lines lie? I mean, if we speak now about the substantive issues on the edge of copyright and AI, like how do we go about data mining? What does it mean to train the systems on existing copyrights? What does it mean in relationship to media companies? Morten, what does it make particularly challenging for tech companies to tackle these substantive issues with potential right holders? in Germany and Europe?

Europe, I think, is challenging because it's not a single issue environment.

If you litigate an AI dispute in Europe, you're really litigating only copyright.

You may be litigating copyright and database rights, but also data protection, Personality rights, platform obligations, unfair competition, consumer protection, all this comes together.

And that makes Europe difficult for companies that want a simple yes or no answer.

Just for example, under EU copyright law, text and data mining is permitted in important circumstances.

But the analysis turns on lawful access.

The purpose of the mining is important.

retention of copies and whether rights have been reserved in an appropriate way.

So under the DSM directive, for example, research organizations and cultural heritage institutions, they benefit from mandatory scientific research TDM exceptions, while the broader TDM exception is subject to rights reservations.

and then we have the european ai act that adds another layer so for generated purpose ai models the ai act requires providers to put in place a policy to comply with union copyright law including respecting rights reservations under the dsm directive and for preparing a sufficiently detailed summary of training content So I think at the same time, Europe is very powerful as a litigation venue because it offers a combination of substantive regulation and practical remedies.

And in Germany in particular, claimants often look for injunctive relief.

And that is important because in AI cases, if you have to stop a conduct very quickly, then obviously that becomes much more valuable than waiting for damages for years.

And so ultimately, I think the other distinctive feature is more rights balancing again.

I think in Europe, disputes are not framed only as property disputes or innovation disputes.

Courts and regulators, they have to consider freedom of expression, as I said before, privacy and data protection, personality rights.

And we have then a lot of different European legislations like the Digital Services Act, for example.

that reflects that balancing exercise for online intermediaries like platforms by requiring a notice and action system.

So there is a lot of different interests involved and that makes it quite complex.

So that gave us a very good tour de raison through all the different rights and aspects and factors that need to be put into the equation for getting the balancing between AI technology and copyright holdings of any type right.