Abstract:

Governing artificial intelligence (AI) is high on the political agenda, but it is still not clear how to define and measure it. We compare four approaches to identifying AI patented inventions that reflect different ways of understanding AI with divergent definitions. Using US patents from 1990-2019, we assess the extent to which each approach qualifies AI as a general purpose technology (GPT) and study patterns of concentration, which both are criteria relevant for regulation. The four approaches overlap on only 1.37% of patents and vary in scale, accounting for shares that range from 3-17% of all US patents in 2019. The smallest set of AI patents in our sample, identified by the latest AI keywords, is most GPT-like with high levels of growth and generality. All four approaches show AI inventions to be concentrated in few firms, confirming worries about competition. Our results suggest that regulation may not be straightforward, as the identification of AI inventions ultimately depends on how AI is defined.

Citation:

Hötte, K., Tarannum, T., Verendel, V. & Bennett, L. (2024). 'Measuring artificial intelligence: A systematic assessment and implications for governance'. INET Oxford Working Paper No. 2024-02.
Download Document (pdf, 1.235 MB)