Artificial intelligence (AI) is a constantly evolving concept, which is struggling to find a consistent definition. Andrew Moore sums it up in one sentence: AI would mean “having computers that behave in a way that until recently was thought to require human intelligence”.

In the economic press or in national and European public strategic documents on AI, we most often find a multitude of associated terminologies such as: algorithm, robotisation, machine learning, Industry 4.0, Big Data, and more generally any name beginning with “e-“, whose readability is not always so obvious.

In this regard, a study carried out in early 2019 by the London-based investment company MMC Ventures states that among the 2,380 European start-ups qualified as “AI”, more than 40% do not use any technology emerging from this field.

In spite of these semantic differences, it is unanimous: AI is the new industrial and societal revolution at work in the world, comparable to the transformations brought by electricity more than a century ago. Geographical competition is raging, and Europe, lagging behind the United States and China, unveiled at the end of 2018 its coordinated plan in the field of artificial intelligence “made in Europe“. It is designed to combine the efforts of the Commission and the Member States to perhaps achieve aggregate GDP targets which, according to McKinsey’s analysts, will be between €2700 and 3600 billion by 2030.

These economic prospects are a dream come true, despite the fact that investment in the European Union is simultaneously low and fragmented. The various European instruments (Horizon 2020, but also private investments) are expected to reach, in 2019 and 2020, €20 billion dedicated to AI.

From 2021 onwards, the same amount should be devoted to the subject on an annual basis. Under the main Community programmes concerned, Horizon Europe (successor to the current framework programme for research and innovation) and Digital Europe (future programme targeting the digital common market), the European share provides for 7 billion euro to be dedicated between 2021 and 2027, to which Member States’ contributions and private investment will be added.

But concretely, in the European AI patchwork, for the SME worried about making a success of its “digital transition”, the urban community in the process of “smart citisation”, the public financer or the investor in search of the next digital “unicorn”, what does artificial intelligence really cover, in terms of opportunities, challenges, change… and above all, what kind of potential financial support?

The digitalisation of the European economy does indeed raise questions in terms of financial strategy. Thus, the amounts stated reveal realities such as a sharp increased in the use of financial instruments to the detriment of the “classic” subsidy model, which requires financial engineering skills that are new for many public and private operators. What support is planned in this area in the myriad tools of the new Community budget programming period?

The uncertainty deriving from the aggregating arrangements of national and Community interventions potentially increases the risk of loss of legibility and therefore of accessibility to the support mechanisms envisaged. Genuine coordination between the actors in charge of AI development, and communication adapted to the different targets represented by… everyone in AI, since the economy, public services and administration, research, training, and citizens will be impacted.

The question of digital skills within companies and the bodies in charge of investment choices also arises, in order to make the ambitions of the European plan compatible with the reality of the current economic and institutional field. Digital Europe is planning only €700 million for the development of “digital skills”. Other funds will be mobilised at regional, national and European level, sometimes at the cost of a detective like or meticulous job to identify the right budget line for a training or education project.

Finally, the European coordinated plan, in its strategy of differentiation with the global AI giants, insists on prioritising societal and ethical issues in the development of AI. These aspects are indeed essential to preserve civil liberties and to understand the challenges of the commercial use of personal data. But in concrete terms, what administrative and legal obstacles or burdens will weigh on the operators on whom the dynamism of the development of innovative and socially responsible projects depends?

Since its emergence, artificial intelligence has been generating and debating these questions. In this period of finalisation of the European budget for 2021-2027 and the definition of the modalities of operation of Community programmes, we can only hope that they will find solutions that take into account the concrete needs of the inventors and users of tomorrow’s AI.

Sandra Ungemach-Benedite