UK offers EU cooperation on artificial intelligence policy and copyright

The UK’s artificial intelligence minister has offered to work with the EU on regulation, and said that London would likely take an approach to copyright in AI that was “reasonably close” to Brussels.

Jonathan Berry took part in a Brussels discussion last week on AI regulation including MEP Axel Voss, rapporteur on the EU’s AI Act, which is now close to being agreed.

While the UK is still not close to enacting similar legislation, and has instead told existing regulators to draw up rules on AI, Berry said the two sides should work with each other and trade information as the technology evolves.

“I think there are really valuable opportunities for shared learning here,” he told the event, held on 15 February, co-hosted by Voss and the UK’s EU mission.

In common with the EU’s AI Office, which is being set up to help member states enforce the act, “the UK’s approach also has a central function, seeking to monitor and evaluate risks, spot regulatory gaps, and ensure cohesive governance. I am certain that we can learn a great deal from each other,” Berry said.

Berry’s comments appear to be part of a broader warming of relations between the EU and UK on science and technology, which have been bedevilled since 2021 by acrimonious, stalled negotiations over UK membership of the EU’s Horizon Europe research programme.

Last September, UK association was finally agreed following a higher-level breakthrough over the status of Northern Ireland in the post-Brexit settlement. 

Since then, relations have warmed, with research commissioner Iliana Ivanova last week visiting the UK to drum up interest among British businesses for the programme.

Science love-in

During his visit to Brussels Berry also met EU research ministers as part of an informal session of the bloc’s competitiveness council. A spokesperson for the UK’s EU mission described his trip as part of a “global science love-in”.

“Areas such as AI are the ones that is probably most critical for us to develop a new relationship on,” said Lindsay Appleby, the UK’s ambassador to the EU, speaking at the AI event. “They're not the kind of areas where you can simply reach for a rule book”.

This cooperation matters because although the EU has written the world’s first comprehensive legislation on AI, the UK is on most measures far ahead of any EU member state in actually creating the technology, and often second only to the US and China. 

Last year, the UK’s AI companies attracted around the same amount of investment as the entire EU. The UK had more AI unicorns, private start-ups valued over $1 billion, than Germany, France and Switzerland put together.

“There is a lot we can do together to advance responsible AI globally,” said Berry. He is open to sharing risk registers with the EU as both map out the potential hazards of AI.

“To be honest, a risk register that only talks about the risks inherent in one country is […] not useless, but it's more useful the more you share it,” he said.

Common copyright

Berry said the UK would also likely take a similar approach to the EU by making sure that creators are renumerated if their work is used in the training of AI models.

Journalists and writers have long complained that big AI companies have hoovered up their work without asking. The New York Times, for example, is currently suing OpenAI and Microsoft for copyright violations. A recent UK House of Lords committee report warned that AI companies were exploiting copyrighted material “without permission or compensation”.

“There's probably a landing zone, reasonably close actually, to what our friends in the EU are doing,” said Berry of the UK’s AI copyright approach.

“It is absolutely crucial anybody coming up with a product of the mind can expect to be rewarded for that in just the same way, regardless of whether AI exists or doesn't exist,” he said.  “And that's so important for our society. That's what makes us who we are.”

Evolving circumstances

Despite promises to cooperate with Brussels, the UK is still taking a more hands-off approach on AI regulation.

Earlier this month, London confirmed it is still not ready to put forward legislation on AI, a reluctance that has annoyed some MPs. “We are going to take our time to get this right – we will legislate when we are confident that it is the right thing to do,” the government said.

Instead of legislation, the UK has given its existing regulators five principles to incorporate into their remits, and insisted that it will remain “strongly pro-innovation.”

Berry described the UK’s approach as “trying to build our system out of Lego rather than out of concrete” that was “adaptive to evolving circumstances with fast moving tech”.

But the challenge the UK faces in trying to set standards for AI regulation is that companies might instead align with EU rules to gain access to a much bigger market.

Asked about this “Brussels effect” in AI policy, Berry said that the creation of AI was currently “truly global”, and it was near impossible to develop a model just in one country.

“So I think they will all have to use the different regulations of the different markets they're in,” he said. “And that's going to be complicated.”