Von der Leyen gives nod to €100 billion ‘CERN for AI’ proposal

Re-elected European Commission President Ursula von der Leyen’s political guidelines addressed calls for huge artificial intelligence (AI) research investments under the “CERN for AI” banner, but proponents and critics say the plan is lacking crucial details.

Several calls for large collaborative AI research projects have employed a “CERN for AI” framing, invoking a similar strategy and goals to the European Organisation for Nuclear Research (CERN), which shifted the centre of gravity of fundamental physics from the US towards Europe after World War II.

Von der Leyen’s invocation of CERN indicates that the Commission is seriously considering proposals calling for €100 billion over five to seven years, more than the Commission’s entire seven-year research budget for Horizon Europe, which is €95.5 billion.

“What von der Leyen said was unspecific enough that it is all going to depend on the details,” Holger Hoos, Humboldt professor for AI at RWTH Aachen University and co-founder of the Confederation of Laboratories for Artificial Intelligence Research in Europe (CLAIRE), told Euractiv.

Hoos said he is “very pleased” that the CERN for AI proposal reached the top level of the Commission.

Von der Leyen’s political guidelines for the next European Commission say that she will propose setting up “a European AI Research Council where we can pool all of our resources, similar to the approach taken with CERN”.

Europe is trailing far behind the US in AI development and computing power, according to recent reports from the European Court of Auditors and the Commission.

Von der Leyen’s proposal of a Research Council suggests it could be part of the Commission’s budget, like the European Research and Innovation Councils, which are funded through the Commission’s Horizon Europe programme, whereas her mention of CERN suggests member state funding.

Proposals floated 

Hoos and CLAIRE has been fronting a CERN for AI idea since 2018,  calling for €100 billion over five years for an AI research project involving a network of research institutes all over Europe with a central facility as a focal point.

In March, the Scientific Advice Mechanism (SAM) to the Commission proposed establishing a European Distributed Institute for AI in Science (EDIRAS), which they described as a “distributed CERN for AI in research”, that should be complemented by dedicated funding through a “European Council for AI in science”.

The Commission has yet to release any official updates on the SAM proposal.

The SAM and CLAIRE proposals overlap, but where CLAIRE calls for focused efforts including a central facility, EDIRAS is fully remote and aims to promote AI in science across the board.

The Centre for European Policy Studies (CEPS) think tank recently issued a report attempting to unify these and similar proposals in a collaboration on trustworthy AI with a price tag of €100 to €120 billion over seven years.

While acknowledging that such a sum is substantial, the CEPS report says “the size of the investment must be assessed against […] the opportunity cost of doing nothing.”

So far, states have focused on national AI initiatives, but Hoos says his impression is that “people are waking up to the reality that if [European countries] don’t do [cutting-edge AI research] together, it will not get done”.

His colleague and co-founder at CLAIRE, Morten Irgens, added it is “more likely to be a strong centre if the Commission gets help from the member states.”

But “somebody needs to take the first step”, he said.

However, not everyone agrees with the price tag for the project.

Though AI is an “extremely important technology”, Professor in Management, Strategy and Innovation and Bruegel Senior Fellow Reinhilde Veugelers is sceptical of the need for a large AI-specific initiative.

“[€100 billion] is a lot of money, […] and without a really good case it would be a lot of money wasted”, she told Euractiv.

Veugelers said the scientific advisory boards are “basically also lobbying for their field”, that AI “is completely new”, and that we should instead “gradually, depending on whether it works or not, increase the budget, because there is no reason for doing this at a large scale immediately”.

Meanwhile, The Information reported in March that Microsoft and OpenAI were drawing up plans for a single gigantic data centre at a cost of up to $100 billion.

In Europe, the chair of the Group of Chief Scientific Advisors, Nicole Grobert, said in a video from SAM: “The European Commission needs to invest, invest, invest”.

[Edited by Zoran Radosavljevic/Eliza Gkritsi]

Read more with Euractiv