The EU is overlooking the need to support investments in Artificial Intelligence, expert says

The European Commission’s strategy on AI initially had two strands: regulation and investment. However, according to Stefano da Empoli, president of the Institute for Competitiveness, the EU has been neglecting the industrial policy part due to a “lack of courage” from the member states.

The I-Com, in short, is a think tank part of the PromethEUs network, with the Elcano Royal Institute in Spain, the Foundation for Economic and Industrial Research in Greece, and the Institute of Public Policy in Portugal.

The network will present a study next week to emphasise the economic potential of Artificial Intelligence at a time when the EU is primarily focused on the regulatory aspect rather than on an industrial policy promoting European companies in this rapidly developing market.

Regulation over innovation

For da Empoli, the focus on investments in this emerging technology that the European Commission had in 2018 has faded out in favour of a purely regulatory approach centred around the AI Act, a landmark EU law regulating AI based on its potential to harm people.

This two-sided approach was still present when the AI bill was first presented in 2021, together with an action plan to coordinate AI programmes and strategies in the EU. However, according to the Italian expert, since then, most of the attention and efforts have gone into finalising the AI regulation, causing an imbalance in the innovation aspect.

“The investment side has been, unfortunately, quite overlooked so far,” he said, pointing out that some parts of the coordinated plan were never respected, such as the yearly monitoring of AI investments.

For da Empoli, the responsibility is more with the member states, which proceeded each in a different direction without real coordination of national AI strategies. Meanwhile, the regulatory track was more straightforward to pursue because it is in the hands of EU institutions.

Regarding the AI Act, the expert stressed the need to balance the containment of risks with innovation while not excessively burdening SMEs and start-ups with prohibitive administrative costs.

Generative AI

Da Empoli pointed out the colossal competitiveness gap between Europe and the United States and other large economies like Canada, the United Kingdom, and Israel in the crucial sector of generative AI, indicated by Europe’s absence of ‘unicorns’ – start-ups that quickly reach $1 billion in market capitalisation.

“The only generative AI company in Europe that could rise to this league any time soon is the French start-up Mistral, but, so far, it was unfortunately peripherical to the generative AI revolution,” da Empoli noted.

“If you look at the investments in terms of venture capital, Europe is lagging. Investments increased across the EU but also in other geographical areas like the US, so the gap was not narrowed.”

Concerning foundation models, da Empoli argued that some rules might also be envisaged in this area but that the EU should not discourage new companies, including European ones, from working on what is currently the most strategic area of AI.

“I see possible side effects that maybe now are not present, but in the future could be an obstacle even for European start-ups.”

International scenarios

The head of I-com noted that on much of EU legislation, like the General Data Protection Regulation and the Digital Markets Act, the EU could count on the so-called ‘Brussels effect’ in setting international standards, but that might not be true for the AI Act.

“The US would, at least, like to be more active, as shown by the executive order signed off by the Biden Administration. Of course, the order is not a legislative act, but maybe even in a polarised Congress, there could be bipartisan consensus next year for legislation,” da Empoli explained.

In his view, the difference from previous digital issues is that the United States perceives AI as a national security matter.

Economic potential

In an upcoming paper, the think tanks detailed the outlook of the AI ecosystem in Southern European countries, notably Italy, Spain, Portugal, and Greece. As most companies there are SMEs, the study argues that public support is needed to ensure businesses embrace innovative AI solutions and integrate them into their processes.

Regarding research and development, da Empoli said EU countries should pool their resources to reach a critical mass of investments in this area, for instance, by introducing the cross-border regulatory sandboxes discussed in the AI Act.

“A real single market for digital and AI applications is the foremost concern for some start-ups and scale-ups. Without it, it is difficult to imagine European companies reaching a scale comparable to the US and Chinese ones,” da Empoli said.

“It’s a dynamic environment. Europe is lagging behind at the moment, but it has the human capital and perhaps even the financial capacity to turn things around.”

[Edited by Zoran Radosavljevic]

Read more with EURACTIV