AI Reduces Wage Growth by 6.7% in Exposed Occupations, Apollo Analysis Finds

AI reduces wage growth more than jobs, says Apollo analysis of labor trends

An analysis by Apollo has revealed that wages in occupations highly exposed to artificial intelligence (AI) have grown 6.7% more slowly after 2023, with no significant impact on employment. This finding, presented by chief economist Torsten Slok, indicates that the primary measurable effect of AI on the labor market has been a reduction in wage growth rather than an increase in unemployment. The study, conducted by Slok and economist Sania Edlich, examined 321 occupations and found that while wages in jobs with high AI exposure lagged behind those with low exposure, employment levels remained statistically unchanged. The wage gap was particularly pronounced among lower-paid workers, with a 10.7% difference in the lowest wage quartile, 5.4% in the second quartile, and 4.0% in the third quartile. No significant effects were identified in the highest wage quartile. The methodology employed in the analysis is noteworthy. The authors matched the 321 occupations to labor statistics data spanning from 2015 to 2025, utilizing the Anthropic Economic Index. This index measures actual AI usage through model interactions rather than merely theoretical exposure. However, the authors acknowledged the limitations of their study. They noted that their exposure measurement is based on data from only one company, and only a fraction of occupations was included in the analysis. Notably, a dramatic figure of 24.3% for service workers is flagged as being based on a small subsample, warranting cautious interpretation. Contrary evidence has emerged, with US statisticians reporting a 0.2% decrease in jobs across 18 occupations significantly exposed to AI, while payrolls overall grew by 0.8%. Additionally, Goldman Sachs has indicated a decline in job openings within fields susceptible to AI substitution, suggesting new entrants to the job market are facing challenges. Diane Gherson, former chief human resources officer at IBM, offers insight into why job losses may not be immediately apparent. She notes that companies might be reducing hires in high-attrition, lower-wage roles without making public announcements about layoffs. Gherson also highlights an accounting distortion that could obscure the reality of workforce reductions. Severance packages can be recorded as one-time restructuring costs, making layoffs appear less severe, while retraining expenses are reflected in ongoing operational costs. As a counterexample, Gherson cites Ikea, which successfully retrained call center employees to become remote interior design advisors after automating much of their previous work. This initiative has reportedly generated approximately €1.3 billion in business. Slok asserts that the economy is becoming more dynamic rather than contracting, pointing to a record high in business formation rates. However, he acknowledges that the productivity benefits of these changes remain unproven, especially as profit margins outside the largest technology firms have not shown significant improvement. Notably, there is a lack of equivalent studies in Europe. The wage trends identified by Apollo may not be visible in most European labor data. Earlier reports by TNW have indicated the impacts of AI on jobs in Europe without assessing wage changes in a similar manner.

Apollo Economist Says AI Affecting Wages More Than Employment | PYMNTS.com

That's according to Torsten Slok, Apollo Global Management's chief economist, who told Bloomberg News on Saturday (Aug. 22) that the impact of AI can be seen on weaker wage growth rather than overall employment. Slok and fellow research Sania Edlich looked at around 300 occupations broken into high- and low-AI-exposure groups and studied how they fared before and after the advent of ChatGPT, finding that the wages of workers with AI exposure grew 6.7% more slowly than people whose jobs had lower exposure to AI. This effect was more pronounced among lower-income workers, the report said. Slok added that AI has accelerated the rate of business formation, which is now at a record high. "So far the dominating effect has been that there is also a much more dynamic economy where people can now invent ideas, use agents, use loops, graphs to come up with ideas and as a result, create more businesses," he said in an interview with Bloomberg. The report said that this analysis provides another data point in a growing debate over how AI will transform the job market. Bloomberg noted that it contrasts with other findings, like a recent Bureau of Labor Statistics report that examined 18 occupations with AI exposure, finding a 0.2% drop in jobs as of May 2025 from the prior year, with overall payrolls climbing 0.8%. And economists at Goldman Sachs said in May that fields "highly exposed to AI substitution" saw a faster decline in job openings than others. Meanwhile, recent research by PYMNTS Intelligence shows the effect of AI on Labor Economy workers, or those earning up to $25 an hour and typically less than $50,000 per year. "The Resilience Deficit: Labor Workers in an Automated Economy" found that AI is reaching all types of workplaces, with 37% of these workers saying their employer had introduced new AI or automation tools in the last 12 months. At the same time, the research also found that lower-income workers are getting less training, showing less confidence and enjoying fewer financial buffers to handle disruption. "The findings also suggested that AI's impact is moving beyond Silicon Valley and corporate offices into warehouses, restaurants, hospitality, logistics and caregiving jobs that make up a large share of everyday consumer spending," PYMNTS wrote earlier this year. For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.