The EU’s effort to regulate artificial intelligence is a risky exercise – Europe cannot afford to lose out in the evolution of AI

Artificial intelligence systems provide game-changing tools for companies, consumers, and public authorities to reap the benefits of digitalisation. The headlines that ChatGPT and large language models have made over the recent months are just a taste of what is to come.

To keep up with the global AI race – that also underpins the digital and green transitions – Europe needs significantly more resources in the innovative development and deployment of AI across all sectors of society. This was duly recognised in the EU’s 2020 white paper on AI which underlined the dual objective of creating ecosystems of excellence and trust around AI in Europe.

Europe needs significantly more resources in the innovative development and deployment of AI across all sectors of society.

Lately, the trust dimension, embodied in the European Commission’s 2021 pitch for an AI Act, has taken a dominant role. The efforts to bolster excellence have been a sideshow. This uneven approach is not tenable.

Companies need less legal uncertainty, not more

More investments in AI need to be unlocked, especially in European manufacturing industries. Here, AI promises to be a key factor in boosting productivity and sustainability by fulfilling the promise offered by connected products and industrial data.    

   

One in five of AI-curious firms in Finland viewed legal uncertainties as a stumbling block to deploying the technology. 

In 2021, according to the official statistics, 17 percent of Finnish industrial companies used AI, with larger enterprises outperforming smaller ones. 19 percent of all companies considered jumping on the wagon. Notably one in five of these AI-curious firms viewed legal uncertainties as a stumbling block to deploying the technology. 

Knowing what has transpired since, particularly in the field of generative AI, it is safe to assume that both the uptake and the curiousness have shot up. Likewise, and less promisingly, the intensified debate about regulating AI or halting its development must have added to the uneasiness many executives and entrepreneurs feel about committing to tools whose rulebook is unpredictable.

It is not too late to correct the course of the AI Act

That said, the EU should ensure that AI systems offered in the European Single Market are ethical, human-centric, and robust. This ought to be done by laying out clearly defined and fully harmonised rules that foster legal certainty and establish unambiguous sharing of responsibilities up and down the AI value chain. They should also be proportionate to make sure that neither national supervisory authorities nor small and medium-sized companies are overburdened by the new obligations.

The EU’s proposal for an AI Act, although welcome, is not without its shortcomings. As the Act finally enters the interinstitutional trilogues this summer, there is reason for concern about several aspects of the text. There is a danger that the Regulation’s key definitions and provisions will turn out to be overly vague and its scope too broad and prescriptive for the purposes of a horizontal legislation that intends to be risk-based and future-proof.

We urge the EU co-legislators to make the necessary course correction and adopt a more balanced touch when finalising the AI Act.
 

DIGITALEUROPE’s recent sandboxing exercise on the AI Act conducted with Europe’s potential tech unicorns points to similar worries, highlighting numerous areas where improvements and further reflection are necessary. The implications are clear: onerous rules will have a chilling effect on the budding European ecosystem of AI excellence. 

In our key messages for the trilogues, we urge the EU co-legislators to make the necessary course correction and adopt a more balanced touch when finalising the AI Act. It is not too late to make the adjustments that would enable a trustworthy and innovation-friendly evolution of AI in Europe. 

For further information:

Joonas Mikkilä, Senior Advisor, Technology Industries of Finland, joonas.mikkila@techind.fi

See also: Key messages for the AI Act trilogues (pdf)