ILO Working Paper - Generative AI and Jobs

This ILO Working Paper titled “Generative AI and Jobs: A Refined Global Index of Occupational Exposure”, provides a valuable and updated analysis of how generative AI technologies may impact jobs across different countries and sectors. Its findings can help us better anticipate changes in the world of work and strengthen our efforts in advocating for fair transitions, upskilling, and worker protections.

This paper extends the survey methods and range of a 2023 paper, notably by asking several GenAI tools (“impartial arbiters”) to extrapolate the human survey findings and to balance out differences in assessments of roles’ exposure to AI by experts and practitioners in the roles themselves. The authors find that, globally, one in four jobs is exposed to some degree to AI technologies, while in high-income countries that rises to one in three. As they note, exposure of roles to AI automation is higher for the female than the male workforce; this difference increases in high-income countries, where women’s and men’s jobs are exposed 9.6% and 3.5%, respectively.

The report’s conclusion stresses that while GenAI tools may substitute tasks, this does not mean that entire roles are necessarily at risk. The authors describe AI’s potential for “unlocking productivity” by complementing human expertise rather than replacing it.

Two observations on the report from our perspective as sindicalist* at the EUI: First, by basing its findings on workers’ own and industry experts’ assessment of roles’ exposure to automation, the report implies that automation is indexed to the wisdom of the workforce. However, as many commentators have noted, role vulnerability comes about not by its actual automatability, but rather its perception as such by a much narrower professional band: those who make decisions on hiring and redundancy. Second, the report’s first annex includes a table of roles graded by supposed vulnerability, of which some at the EUI may want to take note. Lawyers are apparently the most secure in their roles (0.36 in a scale of 0 to 1), while Philosophers, Historians, and Political Scientists are more vulnerable (0.47). In the second-highest risk category are Administrative and Executive Secretaries, and—surprisingly—Economists (0.54 and 0.55, respectively).

You can access the paper here:
Generative AI and Jobs: A Refined Global Index of Occupational Exposure