Tech Commissioner Henna Virkkunen presented the Commission’s Apply AI strategy on Wednesday, outlining a support plan aimed at helping industries across Europe adopt artificial intelligence at scale.
The wide-ranging document outlines a series of lists of measures which it hopes will accelerate take-up of AI. Top of mind is AI development itself: a “Frontier AI Initiative” is set to bring together companies and researchers to unlock “advanced capabilities”. This will include competitions to develop new open AI models, which the strategy says will get free access to EU-run supercomputers.
“The strategy also encourages a Buy European AI approach, especially for the public sector”, Virkkunen said today, and mentioned favouring European innovation through public procurement – though the strategy doesn’t outline specific steps in this direction. A Commission spokesperson could not immediately tell Euractiv which actions Virkkunen was referring to.
The Commission will also open up European compute capabilities to train AI models for defence, and fund “AI for business” masters’ programmes to ensure companies have the skilled workers to tap into the tech.
The strategy outlines support for developing AI models and agents tailored to the manufacturing sector, as well as funding for “acceleration pipelines” – meant to funnel AI from research labs into factories.
In the health sector, AI is envisaged powering advanced screening centres. On climate, an open-source AI model is set to make for better weather forecasts, while – eyeing AI-driven mobility – a group of cities will be asked to volunteer as testbeds for autonomous cars.
Just another strategy?
All of this AI acceleration will be monitored by an “AI Observatory” and discussed in an “Apply AI Alliance” – a refashioning of the existing AI Alliance, a stakeholder talking shop.
But while the strategy is big on announcing broad-brush measures, the finer details – how to actually implement this stuff – are harder to find. The strategy is mum, for example, on what form the planned observatory will take and how exactly it will work.
“Another day, another strategy paper,” said Daniel Abbou, president of the European AI Forum which brings together national AI associations, responding to the Commission announcement. “A lot of what’s in the paper is correct and important. But how should these changes be implemented concretely?”
Abbou points back to the announcement by Commission President Ursula von der Leyen at February’s AI Action Summit in Paris, which included an initiative that was supposed to “mobilise” €200 billion to invest in European AI.
While the Commission is working on establishing large AI model training hubs, called AI Gigafactories, these facilities only account for one-tenth of the teased investment. A spokesperson could not immediately account for the rest of the announced billions.
Measures need money
The Apply AI strategy contains a new funding announcement: €1 billion for AI. This time, at least, it appears to be real money, coming from actual Commission funding programmes.
But private European funding still pales in comparison to investment capital available in the US or China, Emmet King, from AI venture-capital firm J12, told Euractiv. For him, the key to any “AI sovereignty” demands mobilising European savings to invest in powering up innovation.
“While the Commission deserves credit for its proactive approach, the challenge has never been a shortage of strategies – it has been execution,” King also argued, agreeing with Abbou.
He’d like to see all these various strategies and action plans linked to measurable results – like how much AI compute capacity is available in Europe. On that, at least, the Commission might have a concrete answer: in the form of the planned AI Observatory.
(nl, aw)
