Since the ‘AI made in Europe’ strategy launched in February 2020, the US has pulled further ahead. The EU’s problem is a lack of scale and focus. The answer is to adopt CERN’s approach to running large, coordinated and highly ambitious projects
Last month, the President of the European Commission, Ursula von der Leyen, gave her last annual State of the Union address to the European Parliament before the EU elections next year.
It was – naturally - a speech that looked to past achievements as much as future directions - but whatever your view of how much actual progress there has been, van der Leyen cannot be blamed for lack of ambition.
One of her main areas of focus has been to strengthen the EU's technological sovereignty, by promoting innovation and research that will reduce dependency on imported products and services. In pursuit of this, billions have been invested in Horizon Europe, and a dizzying number of technology regulating acts have been put forward.
A crucial pillar in von der Leyen’s vision is a global leadership position in artificial intelligence (AI). In February 2020 she set out a plan, with the aim of the EU attracting €20 billion a year in AI investment over the next decade.
Less than four years later, as AI has started to deliver on the promise of reshaping our economy, our businesses and our lives, it has become clear how important this vision was. But it has also become clear that von der Leyen hasn’t delivered on that vision. In fact, the gap between Europe and the US and China has only grown, despite the fact that Europe has excellent universities and AI researchers. Why is this the case? Simply put: a lack of scale and focus.
The Commission's strategy has been to spend its funds in a way that, in effect, returns them to the member states through a plethora of modest investments, possibly with the hope that large-scale benefits somehow will emerge from the sum of these largely disjointed parts.
Compounding this, these investments are largely uncoordinated with the national programmes of the member states. Unsurprising then, that this strategy has failed to achieve its goals.
The success of US-based AI companies is based on a very different strategy: to concentrate substantial effort and resources on clearly laid out objectives. The breakthroughs have been enabled by sheer scale: scale of investment into large teams of AI experts, vast AI-specific computing resources, and tremendous amounts of data.
Since the European Commission's AI strategy has fallen short in terms of both scale and focus, it is not surprising that four years in, ‘AI made in Europe’ has failed to generate attention, impact, products or value of global significance, and the European AI ecosystem has fallen further behind its competitors.
Europe has spent billions, but the AI infrastructure, algorithms, talents, businesses, research labs and products are stacked higher and better in the US.
How much of a problem is this for Europe?
This is a very serious problem indeed. The failure is not only about AI per se: the impact of applying AI tools across our society and across all sectors of our economy can hardly be overestimated. AI will affect all sectors and all aspects of our lives, because it is poised to play a key enabling role in advancing all fields of science and technology.
In many ways, AI is like the canary in the coal mine – an early indicator of far greater failure to come because by failing to achieve competitiveness in AI, Europe is losing the ability to shape its future and is now quickly becoming deeply dependent on AI technologies developed and controlled by a few foreign companies.
This dependence seriously undermines our sovereignty: economically, culturally and geopolitically. All sectors are affected, along with the public services that form the basis of our society and way of life.
CERN, the European Organization for Nuclear Research established in 1954 in Switzerland, is today one of the largest and most respected centres for scientific research in the world. Scientists go to CERN to work on large, coordinated and highly ambitious projects. When they return home, they lift the scientific ambition, quality and output - not only in particle physics and not only in their own labs: the world wide web was invented at CERN.
In short, CERN is a prime example of the success Europe can achieve if it pursues projects at scale. It is an excellent model for how to effectively boost excellence in AI.
Europe needs a CERN for AI, publicly funded and responsible to the public, not only as a home and focal point for cutting-edge, large-scale AI research "made in Europe", but also to make the benefits of this research accessible broadly across industry and society.
Morten Irgens and Holger Hoos are members of the board of directors of CLAIRE, the Confederation of Laboratories for AI Research in Europe, the world’s largest association of AI research labs.