Europe too ‘defeatist’ on artificial intelligence, says head of European Parliament’s new AI committee

Europe can claw its way into the artificial intelligence (AI) race but it has to get tougher and savvier, according to Dragoș Tudorache, the Romanian MEP recently installed as the chair of the European Parliament’s new committee on AI.

“It is commonplace to lament our place in the AI race. Clearly, we are behind in investment. By quite a sizeable margin too,” Tudorache said. “But I do believe if we wake up soon enough, we can make up this gap.”

The message for member states negotiating the next EU long-term budget is, “Ambition without money doesn’t work,” he said. “If the COVID crisis has brought us anything, it’s the understanding that we need a serious digital push.”

However, that’s not what Tudorache, who sits with the liberal Renew faction in parliament, is seeing. Europe suffers from an AI talent shortage that is exacerbated by slow efforts to reform immigration policies.

Tudorache admires the cunning of the Canadian government, which has paid for billboards in Silicon Valley that pointedly aim to lure AI talent with the promise of easier visa conditions.

“We have not figured out a way to attract talent from outside the EU. We’ve been battling on a blue card [skills] visa, but have been politically blocked for years,” he said. “We’re all too obsessed with migration in Europe. You can’t win the race if you don’t have the right talent. We’re not doing enough to make models to attract talent that may be running away from [US President] Trump and the like.”

Keeping tech talent is another challenge for Europe. “We have great mathematics schools, lots of home-grown potential. We’re losing people, mostly to the US,” Tudorache said.

Europe is not among the top competitors in AI, according to Stanford University’s AI index report, which found the US and China account for almost all private AI investment in the world.

The European country with the most meaningful presence on AI is the UK, which has left the EU and has hinted that it may detach itself from EU data protection laws in the future.

Tudorache is calling for more money in the EU’s next seven-year budget to boost AI.

“We need beacons of research and innovation – efforts to make us more recognisable - and sexy, even, to the world. For many, this is important. If you have a top-notch research centres, with the right panache, the kind of places you can boast about, then [you] may consider moving to Europe,” he said.

New AI law

The EU will next year attempt to establish first-of-a-kind rules on the ways that companies can use AI. Supporters of regulation say proper human oversight is needed for a technology that presents undeniable risks. Others warn that the new rules could stifle innovation, and set the EU back in its attempt to become a true AI player.

Those who support legislation are concerned it will take too long to regulate the sectors where it is deployed.

“It’s going to be much faster than GDPR (general data protection regulation),” Tudorache said. “I don’t think the first wave of legislation will be as complex as GDPR.” The EU’s data privacy rules took around nine years to arrive on statute books.

“It’s like building a rocket, you advance stage by stage. We should not be rushing to legislate all aspects of something that is still in early existence. The key word is differentiation: to have a risk-based approach,” said Tudorache.

The parliament’s committee on AI, created in July, will study the impact and challenges of rolling out the technology. The committee’s ranks include Romania’s Dan Nica, Portugal’s Maria da Graça Carvalho and Spain’s Pilar del Castillo Vera, three members who sit on ITRE, the committee on industry, research and energy. Also on the new committee is Andrus Ansip, former European Commission vice president, and former holder of the digital single market portfolio.

Tudorache talks about creating a “risk register” for different AI applications. In February, the European Commission presented its AI white paper, which states that new technologies in critical sectors should be subject to legislation. It likened the current situation to "the Wild West" and said it would focus on "high-risk" cases.

What AI does he consider risky? “There’s a broad consensus that risky includes everything that has an impact on human life, whether it’s decisions on health treatment or intervention, applications that involve credit qualifications or insurance ratings. Anything that involves the rights of individuals, in other words, would be considered a high risk,” he said. “You wouldn’t allow an algorithm to take a decision from A-Z.”

One idea floated, and then dropped, by the European Commission earlier this year was a temporary ban on the use of facial recognition in public areas for up to five years. There are grave concerns about the technology, which uses surveillance cameras, computer vision, and predictive imaging to keep tabs on large groups of people.

Would he support a temporary ban on facial recognition? “Again, any regulation needs to be differentiated. Facial recognition may be used in certain applications that are not invasive [for the general population],” Tudorache said.

“I’m not excluding a ban. In any case, facial recognition should be on the top shelf of risk when it comes to regulation,” he said.     

A recent investigation by human rights campaign group Amnesty International, said several European tech companies have been selling surveillance technology to Chinese security agencies.

Tudorache said in the wrong hands, AI tech is potentially authoritarian. “Clearly there are elements of the [Chinese] model, which are not compatible with our model. We believe very strongly in the rights of individuals. We’re not very soon going to see eye to eye with China on this issue,” he said.

Defence is the other big, controversial area for AI applications. Some are pushing for the EU to ban the development of lethal autonomous weapons. “The risks are very big. No decisions should be taken solely by a machine without any human control,” Tudorache said.

Tudorache says officials will also assess whether the EU’s prioritising of data protection, through GDPR, has harmed AI growth in Europe.

“There’s an evaluation ongoing. I have always said I would like it to look at the impact on business. The smaller you are, the more costly it is to comply [with GDPR]. I need to see some figures but my instinct tells me the impact exists,” he said.

‘Defeatist’ EU?

Tudorache challenges the prevailing view that Europe missed the boat on building giant internet platforms, and now has to search elsewhere to find its niches.

This line of thinking says that it’s too late to build Facebook or Google-equivalents, using mountains of personal data, and now the fight has to shift to exploiting industrial datasets created by digitised machinery.

“There was quite a defeatist sentence in the recent EU state of the union speech from [commission president] Ursula von der Leyen which said, we’ve missed the wave, full stop. No big players are European, that’s true. But we don’t have to take that as a fait accompli, a given.”

“Personal data is something we all produce, every second. Right now, our data is feeding algorithms somewhere, and not in Europe. But we are, each of us, every second generating the next wave of data.”

“I would like to see a level of ambition on personal data we continue to produce. What is the model of management for this data from now onwards?”

For Tudorache, the story doesn’t end with hiking up tax on the big American tech giants. “That’s not enough. The discussion has be more about making our own models,” he said.