The EU position in the AI Global Race and the situation in the Member States

Article
Maria Rosaria Della Porta

The interest in AI has grown dramatically in Europe. According to some estimates, the European AI software market alone should see a significant growth in the coming years, with revenues increasing from around US$ 2.09 billion in 2018 to an expected US$ 26.5 by 2025.

Therefore, the European Union has become one of the most important AI players in the world, together with the United States and China. All three countries have adopted strategies for the development, implementation and regulation of AI. In Europe, the EU Commission has promoted an approach that is dubbed as “ethical and trustworthy AI”, fostering cooperation between the public and private sectors, as well as ensuring a set of key values (e.g., transparency, accuracy, robustness and non-discrimination).

A COMPARISON OF AI DEVELOPMENT IN CHINA, THE US, AND THE EU

However, in comparing the EU to China and US, a pattern of a clear competitive disadvantage seems to emerge. This especially concerns the gap in the overall amount of investments made, the main reason for Europe lagging behind. According to Stanford University estimates, in 2020, the US continued to hold a dominant position in private AI investment, with US companies investing $23.6 billion in AI, compared to China’s $9.9 billion. Instead, the level of European private investments was seen to be much lower, amounting to only $2 billion.

Moreover, the US and China account for most of the AI start-up investments. Together, the US and China made up over 80% of the monetary value of VC investments in AI start-ups, while the EU-27 represented almost 5% Only the UK start-ups represented a little more than 4% of the total value.

Thus, it is clear that the EU is lagging behind, and this gap is also seen in the field of AI research. For example, in terms of patents, the EU-27 with about 3,000 AI patents registered in 2021 ranks after the US with almost four times as many patents registered.

I-COM 2021 INDEX ON AI DEVELOPMENT IN THE EU COUNTRIES

Analysing the situation at the level of individual Member States, substantial differences can be found, with some countries even managing to keep pace at international level, while others appear to be unwilling to fully adopt AI. This situation certainly does not favour the UE’s position in competing in the global race for AI.

In order to provide an idea of the degree of AI development in European countries, I-Com has developed a new index, presented during the I-Com Symposium “EU’s Path to Competitiveness. How Digital, Energy and Health can L.ead the Way Forward”. This index takes into account some variables relating to the industrial and research AI ecosystem in the various Member States as well as the level of the adoption of some AI technologies. These include:

  • number of AI firms (per capita value)
  • number of AI research institutes (per capita value)
  • AI patent applications (per capita value)
  • enterprises using 3D printing (in % of enterprises)
  • enterprises using service robots (in % of enterprises)
  • enterprises using industrial robots (in % of enterprises)
  • enterprises analysing big data internally using machine learning (in % of enterprises)
  • enterprises analysing big data internally using natural language processing, natural language generation or speech recognition (in % of enterprises)
  • enterprises with a chat service where a chat bot or a virtual agent replies to customers (in % of enterprises).

From I-Com’s analysis it emerges that Ireland tops the rankings with a score of 100, followed by Malta and Finland, with 95 and 78, respectively. Despite being smaller countries – slightly exceeding 11 million inhabitants in total – these states have a good Al ecosystem. Ireland is particularly emerging as a leading player with 273 AI firms, many having filed patent applications. Moreover, the percentage of Irish enterprises analysing big data internally using machine learning (20%) is much higher than the EU average (2%). Malta and Finland also have much higher percentages of enterprises than the European average where using machine learning for big data analysis and using chat bots or a virtual agent to respond to customers is concerned. Germany (with a score of 44) and France (38) follow at a distance. The two main EU countries in terms of AI firms are way behind because, for size, their industrial and research AI ecosystems are not yet fully developed and the use of AI technologies by total enterprises is very low.

At the bottom of the ranking, we find the countries of Eastern Europe, where both the industrial and research AI ecosystem sees a lower number of active AI players or where the level of adoption of technologies is very low.

CONCLUSIONS

In the coming years, many segments, if not all, of the economy and daily life will probably be affected by the development of AI technologies and their applications. Therefore, it is of crucial importance that the EU, MSs and stakeholders act fast to boost the competitiveness of the European AI market. A key element to guarantee the full development of AI in Europe will be a close collaboration between the private and the public sectors. It is also essential to strengthen the field of AI research and public and private investments as well as digital skills.

Moreover, the proposal of an Artificial Intelligence Act is held to be of key importance especially in establishing a solid framework for the protection of European citizens and the development of trustworthy AI technologies. Finding the right balance between developing AI technologies and promoting the model of trustworthy AI appears to be the only way towards creating a legal framework that will be able to respond to the speed of this fast-growing technology.