[1/2] Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence (AI column)

Interesting. study On the impact of AI on employment from students from Stanford University: Erik Brynjolfsson, Bharat Chandar and Ruyu Chen. By the way, Brynjolfson is the co-author of the cool book "Machine, Platform, Crowd," which I'm already talking about. told. The purpose of this new study was to assess recent changes. (2022 - 2025 years) in the U.S. labor market in the professions most affected by generative artificial intelligence. Data for analysis were obtained in cooperation with the company ADP The largest provider of payroll software in the United States, providing authors with access to extensive, up-to-date employment information. The executive summary of the study looks like this

1. Reducing the employment of young professionals in jobs with high AI impact. Since the widespread introduction of generative AI (ending 2022 year) employment of young workers 22–25 Over the years, the most “AI-subject” professions have decreased significantly. (13percentage-decrease)Although the employment of older professionals has not fallen. 2. Overall employment growth with stagnation for youth in vulnerable areas. Overall employment in the economy continues to grow steadily. In other words, there is a redistribution of jobs: in general, jobs are added, but growth does not affect young workers in highly automated roles. 3. Greater reduction is where AI automates rather than helps. Employment declines are concentrated in jobs where AI is more likely to automate than complement human labor. In professions that use AI as a supplementary tool (reinforcement) Human work, no serious reductions were found. However, in tasks that AI is able to fully automate, the employment of young people decreased most dramatically, for example, in software development or customer support services. 4. The effect persists even with company/industry factors in mind. The authors tested whether the observed reductions were due to any specific shocks at the level of individual firms or industries. Turns out that's not the point. 5. Adjustments are made through layoffs, not salaries. Analysis of the labor market shows that adaptation to AI occurs mainly due to changes in the level of employment, and not through changes in the level of wages. 6. The results are stable in various tests and samples Authors tested the impact confounderBut it turns out it wasn't about them. All of the above trends have continued under various additional checks. For example, excluding tech companies from the analysis, as well as occupations that easily switch to remote work, did not change the general conclusions.

Taken together, these findings provide the first large-scale evidence that the AI revolution has already begun to visibly and disproportionately affect entry-level workers in the U.S. labor market. Young professionals in high-tech and automated professions act here as “canaries in the mine” – an early indicator of the coming changes.

In post I'll talk a little bit more about the methodology of this study, and I'll also share the authors' thoughts on the longer-term effects of current changes.

#AI #Software #Engineering #Management #Leadership #Data #Whitepaper #Metrics