We estimate that if the pace of automation adoption is in the midpoint of our range of scenarios, about 15% of the global workforce, or 400 million workers, will be displaced by 2030 (pdf). (That range stretches from almost nobody, if automation adoption is slow, to 800 million, in the event of very rapid automation).
That’s a lot of people. But he maintains we don’t have to fear being pushed out of the workforce:
At the same time, as new technologies and the productivity gains they bring generate additional labor demand, many jobs will be created; based on historical precedent, we expect 8 to 9% of 2030’s labor supply will be in new jobs we cannot yet foresee. On top of that, we have estimated that between 555 million and 890 million jobs could be created by 2030 from select catalysts. This includes rising incomes (especially in emerging economies), which generates spending and higher consumption, which in turn creates additional labor demand to supply consumers with those products.
Overall, we expect there will be enough work to ensure full employment, with the gains offsetting the losses under most scenarios. But the transition will be jarring.
This somehow doesn’t gibe with others’ perceptions or projections about the impact of AI and automation. And if the ‘additional labor demand’ is increasingly being provided by robots and AI, the logic collapses.
He argues for more STEM and tech skills, but he fails to address the fundamental problem: if you look at this table of occupations being automated, show me examples of successful retraining of manufacturing workers and farmers in high-skill science and programming jobs?