This dimension of this study addresses societal aspects of AI. In order to fully consider the digital transformation process, the discussion on economic aspects (i.e., industry, investments made in AI, and research and developments) must be completed along with societal considerations.
Indeed, it is fundamental to have insights regarding the way the opportunities opened by AI are enjoyed by society, especially with respect to who develops AI and who will work with it in the near future. Thus the AI Watch Index includes seven indicators to target the level of diversity among active researchers in the field, and to investigate how the educational offer related to AI content is distributed among Member States.
The consideration of diversity among the AI research community is relevant, as there are many concerns about the ethical issues related to AI, and therefore it is advisable that the research communityshouldpresent heterogeneous characteristics in order to reduce bias-related risks. Indeed, the presence of diversity should encourage the consideration of multiple perspectives in the development process. And this, in turn, should make the technology fairer and more neutral. At the same time, diversity in research may be the result of inclusion policies which are worth monitoring.
The second aspect considered to address the theme of AI and society, that is academic offerings about AI content, provides elements to proxy the supply of AI skills of future cohorts of workers, which will affect both the individual employment opportunities and the overall human capital present in the economic system to support the innovativeness of industry. In fact, it is likely that the future availability of skilled workforce will have positive consequences on the overall competitiveness of the economy. Also, as AI competences become increasingly desirable in the labour market, they will constitute an increasingly relevant part of a worker’s skill-set, and having them will likely have consequences on families’ incomes and individual well-being. Finally, this technology is so pervasive that having a basic knowledge of its elementary principles is going to be useful for several aspects of one’s personal life (e.g., security, privacy).
Regarding the diversity of the active AI research community, the index considers four indicators of the diversity of participants in a set of international AI conferences. Even if all these indicators present upward trends, translated into an increase of gender, geographical and business-academia diversity, still their values are not close to 1 – meaning maximum diversity. In other words, although in general terms heterogeneity has improved among AI researchers, there is still room to improve diversity in AI teams.
The indicators related to education show that AI content intensity in official studies is heterogeneous across EU Member States, as drawn from a selection of AI programmes taught in English language. Some have a low proportion of university programmes with AI content, for example Slovenia, Luxembourg, Croatia and Bulgaria. They have very low numbers of AI programmes in the total offer of bachelor’s degrees, and also small numbers of available places for students in programmes that contain any type of AI content. Germany is the country with the largest number of available university places with AI content in both master’s and bachelor’s degrees. Other Member States have much less availability of places, with a few positive exceptions, such as Poland and Romania, with regard to bachelor’s degree-level courses, and France and Italy in the availability of places in AI-related master’s degrees.
For most Member States, the presence of AI content in master programmes is higher than in bachelor programmes. The same pattern is detected for the AI intensity in university places – or proportion of university places including AI content – which shows larger percentages in master’s degree-level programmes for almost all Member States. This seems to indicate that AI is considered to be specialised content proposed mostly in a phase of the education path at which basic knowledge has already been provided to students. Indeed, this reflects the characteristics of AI, an advanced technological domain. At the same time, its pervasiveness in the daily life and in many aspects of society and economy should also encourage a wider provision of related contents in less advanced courses (e.g., bachelor’s degrees).
AI Societal aspects indicators
S1-S4: Diversity in research: Gender diversity index, Geographic diversity index, Business diversity index, Conference diversity index
The Societal Aspects dimension includes four indicators addressing Diversity in Research. They measure different aspects related to diversity among participants (i.e., keynote speakers, conference organisers and authors of papers) at a set of international AI conferences.
S5: AI in university programmes in the EU
This indicator evaluates the intensity with which AI is considered in official curricula, as a proxy of AI skills acquired by current students (and therefore, future workers).
S6: University places with AI content in the EU
This indicator measures the number of available places in university programmes with AI content for each Member State.
S7: AI intensity in university places in the EU
This indicator measures the proportion of available places in university programmes with AI content in total number of places in university programmes by Member State.