John Whelan: Will the EU's penchant for over-regulating stymie the AI miracle?

European Commission president Ursula von der Leyen has said she wanted the future of AI to be made in Europe.

For believers, this is the opening act of a decades-long boom. 

Certainly, Goldman Sachs is a believer, estimating AI could add as much as $20trillion to the US economy over the next decade.

There is no such prediction for AI’s impact in the EU, but last week at the launch of two new AI strategies, Ursula von der Leyen, president of the European Commission, said she wanted the future of AI to be made in Europe.

Sceptics are lining up, predicting that AI’s trillion-dollar moment could end like the dot-com decade. But that won’t stop the euphoria. 

At last week’s Investor Day at Dogpatch in Dublin, the euphoria was evident amongst the 20 companies showcasing and pitching for funding for their AI-designed businesses.

However, whether the two new EU strategies will assist or obstruct start-ups or the more established businesses in adopting AI into their operations remains to be seen.

The first strategy, the ‘apply AI’ strategy, includes sector-specific commitments for industries such as health, manufacturing, defence, and energy.

Measures to support compliance with the EU AI Act

General actions include measures to support compliance with the EU AI Act by default and by design, in respect of relevant AI models and AI systems.

In this regard, the commission has set up a new AI Act Service Desk, which it said businesses can use to “access all relevant information about the AI Act, navigate its content, understand how it applies, and get tailormade answers to any question related to its implementation”.

Businesses will also be able to use the hub to check “whether they are subject to legal obligations and understand the steps they need to take to comply”.

This sounds like more of the usual commission’s over-regulating approach and is highly unlikely to appeal to entrepreneurs or busy company managers.

“Why is it so difficult to get AI done in Europe? Simply because we started with regulating, to keep AI under the thumb,” Europe’s leading tech company ASML’s chief financial officer Roger Dassen told a Dutch business grouping earlier in the month.

The commission’s new strategies contained no announcement in relation to whether it will delay the implementation of certain AI Act provisions, amidst calls from industry both in Europe and in the US to do so. 

Only some of the chapters have taken effect so far; prohibitions on certain types and uses of AI began applying in February, while rules impacting providers of so-called ‘general purpose AI models’ came into effect on August 2. 

Rules for ‘high-risk’ AI systems in 2026

Rules applicable to ‘high-risk’ AI systems do not come into effect until August next year, as plans stand right now.

The second strategy paper, the ‘AI in science strategy’, lays out plans to create a “virtual institute that pools excellent talent, compute, data and research funding for AI”. 

This is to be called Resource for AI Science in Europe (Raise). Under the Raise umbrella, it is envisaged that European scientists will “advance AI technologies and apply them to the toughest scientific and technological challenges” and have access to EU funding, data, and computing power. 

The Raise initiative is essential to help the EU compete with countries like the US, China, Japan, and the UK on integrating AI into scientific work, according to the commission.

As part of the launch, the commission confirmed the setup of ‘AI Factories Antennas’ in seven member states, including Ireland. 

These antennas will collaborate with AI supercomputer factories to provide AI communities in each country with access to world-class AI-optimised supercomputing resources. 

This is likely to be the main attraction for businesses, as access to the high-speed data analysis provided by supercomputers is the key to creating meaningful AI models that companies can usefully use in their specific businesses. 

Just 13.5% of EU businesses use AI currently, according to the commission.

The commission and member states seem to be aware that the EU is well behind AI developments in the US and China, and agreed to boost investment in the sector by maximising resources of both the Horizon Europe and Digital Europe programmes, which will make €134bn available for digital and AI development. 

The commission expects industry to top this investment figure up to €200m over the next decade. In the process, the current 250 Digital Hubs across member states will be refocused as ‘experience centres’ for AI.

But will the penchant for over-regulating stymie the AI miracle in the EU?