Introduction
The G1 – AI economic players indicator measures the number of the three types of economic players included in the AI landscape analysis:
- research institutes (including universities),
- firms and
- governmental institutions.
The breakdown of AI players by organisation type makes possible further analysis of the relationships between research, industry and government in different geographic areas and facilitates the assessment of different properties of the entire AI landscape and local areas.
The types of AI activities tracked are: business activities (firms with a core business in AI), research activities (scientific articles in top AI and robotics international conferences), and innovative activities (AI-related filed priority patent applications). Additionally, a fourth group is considered: AI-related EC-funded projects. International comparability is given when the first three groups are used. EC-funded projects are only used for the in-depth analysis of the EU and its Member States.
The second indicator, G2 – AI economic activity intensity, expresses the presence of AI economic players in relation to the size of the economy. It is calculated as the ratio between the number of AI players and GDP in billions of euro. Hence the G2 indicator makes possible a comparison of the size of the AI landscape for each geographic area.
Worldwide AI landscape
The US is home to the largest number of AI economic players, with 13,770 organisations. It is followed by China, which has 11,362 players, and in third place is the EU, with 5,933. These three dominate the worldwide AI landscape. However, we can see a clear distance between the US and the EU, since the former has more than twice as many players as the EU.
Nevertheless, the number of AI players has limitations as an indicator, as it does not consider the size of the country’s economy, the size of the players themselves (e.g., a small start-up and a multi-national both each count as one player) or their intensity of AI engagement (e.g., number of AI publications or AI patents). The ratio of number of AI players to GDP provides a different perspective of the AI landscape, giving a relative measure that enables one to compare countries of different economic weight. The UK holds simultaneously a remarkable number of AI players and the highest ratio of players to GDP of all geographic areas considered. This highlights the consequent loss of strength of the EU AI landscape in the global perspective after the Brexit. While the UK has 0.98 AI players per billion euro of GDP, in the US the ratio is 0.79, in Canada 0.78, in China the ratio is 0.62 and in the EU 0.34. Therefore, even if the US and China have the highest number of AI players in the AI landscape, is the UK the country with the highest AI player intensity.
The majority of AI players are firms, followed by research institutions and government. While firms account for more than 85% of the total number of players for all the cases, for some Asian countries the participation of research institutes is higher than the global average (6.5%). This is the case in China, where 12.6% of the players are research institutes, Japan (12.4%) and South Korea (14.2%). Even if this information does not assess the quantity or the quality of the research that is carried out, the presence of research institutes in the AI landscape is very important, due to the obvious link between research and innovative outcomes. The position of the EU in this respect (6.0%, without considering the impact of EU-funded projects) is halfway between the very poor presence of research institutes in the US (2.6%) and the UK (2.1%), and the higher proportion of research institutes found in China, Japan and South Korea. Government institutions play a minor role in the AI landscape. No government participation is observed in Africa, Canada, India and the Middle East.
EU AI landscape
Within the EU, Germany and France are the two countries with the highest number of AI players, with 1,136 and 1,055 players, respectively. They are followed by Spain, which has a much smaller number of players (614), while the rest of the Member States do not have more than 450 AI players. Estonia, with 66 AI players and a ratio of 1.57 AI players per billion euro of GDP, is the Member State with highest AI player intensity. The ratio in Malta is also remarkable, 1.02 AI players per billion euro of GDP. For the rest of the Member States this value is smaller than 1, indicating that there is less than one AI player per billion euro of GDP.
This graph provides a view on the composition of AI players in the EU by type of organisation. Firms are the predominant type of AI player in all Member States, and governmental institutions account for only a small proportion. The presence of research institutes is significantly high in Romania (16.05%), Greece (14.82%), Slovakia (14.29%), Slovenia (10.34%) and Italy (9.67%).
