After much brainstorming, the EU has unveiled how it plans to catch the US and China when it comes to artificial intelligence (AI).
Ursula von der Leyen, announcing the new strategy, said AI could be hugely lucrative but that it came with key risks.
Here we take a look at what Brussels wants to do, why and what critics think of the plan.
What is artificial intelligence?
Artificial intelligence sees machines use intelligence typically associated with humans, such as learning from mistakes and adjusting to new inputs.
Examples include self-driving cars, voice-powered personal assistants like Amazon's Alexa and the latest e-mail spam filters.
Why is the EU acting now?
"AI is developing fast, which is why Europe needs to maintain and increase its level of investment," said the European Commission in a press briefing. "At the same time, AI entails a number of potential risks that need to be addressed."
The risks include concerns around citizens' privacy, human rights and livelihood.
But the benefits are huge, according to European Commission President Von der Leyen.
"It is €300 billion right away, this is 2.4% of the Gross Domestic Product – this is today. In five years, it will be estimated three times as much," she said. "Today, we have 5.7 million jobs, in five years, these will – estimated – be about 11 million jobs."
There are also fears Europe is lagging behind other parts of the world.
Data recently published by the World Intellectual Property Organisation showed that Chinese and US firms account for about 85% of AI-related patents globally.
What's in the strategy?
The EU says it wants to develop a "framework for trustworthy artificial intelligence", focusing on three key objectives:
- Technology that works for people;
- A fair and competitive economy; and
- An open, democratic and sustainable society.
Practically, authorities should be able to test and certify the data used by the algorithms that power artificial intelligence in the same way they check cosmetics, cars and toys, the bloc said.
The focus for such preliminary certification is on high-risk systems rather than on all AI "in order to avoid unnecessary burden for companies to innovate", the strategy said.
Examples of high-risk sectors include medical equipment, automated driving, or decisions on social security payments.
The EU also wants to use unbiased data to train high-risk artificial intelligence systems so they can avoid discrimination.
EU leaders said they wanted to open a debate on when to allow facial recognition in remote identification systems, which are used to scan crowds to check people's faces to those on a database.
It's considered the "most intrusive form" of the technology and is prohibited in the EU except in special cases.
The EU strategy includes an ambitious investment plan of €20 billion per year over the next decade, against €3.2 billion in 2016.
Watch: Von der Leyen unveil the EU's strategy on artificial intelligence
How much support do the proposals have?
Initial reactions to the proposals in Europe were overall positive.
Interviewed on Euronews Now, MEP Eva Maydell, of the European People's Part,y hailed the strategy as "future-oriented".
She said it would pave the way for European companies to be more competitive.
Andrea Renda, head of regulatory policy at CEPS, a think-tank, told Euronews that he was "very enthusiastic" about the EU's new proposals and their ambition.
"Part of this package is to ensure that Europe is fit for the digital age, and part of it is that the digital age becomes fit for Europe," the expert said.
This includes ensuring the development of AI technology becomes aligned with fundamental rights and European values, he added.
Digital Europe, a trade association representing the digital industry, welcomed the new strategy in a statement.
"Rather than rush to regulate emerging technologies, we are pleased to see a commitment to work with industry and civil society and enhance the EU’s robust legal framework where there are gaps. It will be important to keep new regulation focused and limited to truly high-risk cases," said Digital Europe Director General Cecilia Bonefeld-Dahl.
What do critics say?
Some actors in the industry are worried that it will be difficult to define the nature and scope of high-risk cases. They also fear that the kind of preliminary certification envisioned by the EU will make the process more complex and bureaucratic.
The new regulations will have far-reaching implications for US tech giants such as Facebook, Google or Apple.
"American companies, large cloud operators, will probably see their potential future market share shrink a little bit as a result of these proposals," Renda noted.
On Monday, Facebook CEO Mark Zuckerberg met top European Union officials as the company warned that potential regulation risked stifling innovation.
However, In an op-ed published in the Financial Times, Zuckerberg said big tech companies such as Facebook need closer government supervision.
"I believe good regulation may hurt Facebook's business in the near term but it will be better for everyone, including us, over the long term,'' he wrote. He said new rules should be clear and balanced and it shouldn't be left up to individual companies to set their own standards.
On the consumers' side, critics say the strategy could go even further to protect rights and privacy.
“Too much data is currently concentrated in the hands of a few industry players who use it exclusively for their benefit,” said Monique Goyens, Director General of The European Consumer Organisation (BEUC)
“It is good that the EU wants to legislate how data can be used better but when it comes to personal data, it must always be the consumer to decide whether their data is collected and how it is shared.”
How do EU policies compare to other global players?
The EU has become a major global player in tech regulation. It has already pioneered strict data privacy rules and issued multibillion-dollar antitrust fines against the likes of Google.
Some of the bloc's new proposals, such as preliminary certification for high-risk AI, are completely novel.
"It is the first time at the global level that a policy-maker comes up with such a comprehensive framework for constraining and steering the development of AI technology," Renda said, adding that the policy could become a "reference for policy-makers all over the world."
The EU focus on regulations comes in sharp contrast with the US approach, which is based on flexible guidelines rather than binding rules.
Another major challenge for the bloc is regulatory fragmentation - tech companies trying to access the EU market face a myriad of different national regulations, unlike the US or Chinese markets. Digital actors hope the bloc's new strategy will offer an opportunity to tackle the issue.
The EU strategy is open for public consultations until May 19 and will be fine-tuned based on feedback. An impact assessment will also be conducted.
"The Commission has not rushed its own thinking for a legislative proposal within the first 100 days," Renda told Euronews.
"This White Paper enables us to have a period of reflection, to continue discussing where we want to go in the next five, 10, 15 yeas," Maydell said.
Further regulations will be introduced later this year, the Commission said, including "a Digital Services Act to establish clear rules for all businesses to access the Single Market, to strengthen the responsibility of online platforms and to protect fundamental rights."