Introduction
The R1 – AI players in AI R&D indicator shows the presence of players active in AI R&D activities, to assess how distinct geographical areas are involved in the technical development of AI. The indicator further distinguishes between the type of organisations involved: firms, research institutes (including universities) and governmental institutions.
World
This indicator offers insights about the players taking part in AI-related R&D activities. As expected, the presence of research institutes is more concentrated in frontier research (proxied by participation in international AI conferences), while firms opt mainly for the development of patents. Nonetheless, some elements deserve to be discussed.
Firstly, the US leads in terms of percentage of firms involved in frontier research. Considering also the overall size and role of the US in AI, this element may suggest that the competitiveness of this country will last, as the private sector is remarkably involved in the scientific domain of AI. Indeed, a considerable presence of firms in frontier research activity should guarantee easier flows of knowledge to be transferred from the early phase of scientific research to the final phase, in which the patented innovations are incorporated in industrial processes. This could explain why the US is the geographic area that has not only the highest percentage of firms developing frontier research, but also one of the highest percentages of firms that file patent applications.
Secondly, other two geographical areas show considerably higher percentages of research institutes active in patents: other Asian countries (which include, for example, Taiwan, Malaysia and Singapore) and, in particular, other European countries (which include, for example, Russia, Switzerland and Norway). These substantially higher values suggest that in these areas the activity of research institutes is more directed towards the development of applied innovations.
The European Union
The R1 – AI players in AI R&D indicator at the EU Member States level also includes the R&D activities carried out in the context of EC-funded projects. We can see a very different composition of AI R&D players by organisation type based on the different types of R&D activities considered.
Firstly, when looking at the first graph below showing frontier research activities, most players in all Member States are research institutes. For some Member States, such as Spain, Slovenia, Portugal and Poland, research institutes are even the only type of players involved in this kind of R&D activity. Governmental institutions have no representation in any of the countries except for France. Belgium and Czechia stand out as the two Member States with the highest participation of firms in frontier research.
The second graph shows the composition by organisation type of players filing patent applications. As expected, firms constitute the main type for all countries, only Luxembourg and Slovakia have a stronger participation of research institutes than firms in patenting activities.
Finally, in the third graph we find that the composition by type of player participating in EC-funded projects is again dominated by firms in all Member States. The distinctive feature seems to be the higher participation of governmental institutions, whose participation in the previous two types of R&D activities is anecdotal. Nevertheless, governmental institutions remain in a third place by order of importance when compared with firms and research institutes, including in EC-funded projects.
Complementary to the analysis by organisation type of AI R&D players, this next graph shows the absolute number of AI R&D players by type of R&D activity. It is noteworthy that for activities related to EC-funded projects the number of players is considerably higher than the number of players involved in patents or scientific publications.
Therefore public funds significantly support the AI domain in the EU. This may, on the one hand, lead to further growth of the AI sector, but, on the other hand, may be interpreted a signal that the EU AI landscape relies excessively on public support.
In fact, a deeper analysis of AI players participating in EC-funded projects reveals that for many of them this activity is the only AI-related activity they carry out, as they do not have a core business in AI nor any other AI-relevant activity, such as patent applications or frontier research (Righi et al., 2021).
