Economists Torsten Slok and Sanya Edlich of Apollo Global Management LLC released research on Thursday, adding new evidence on the impact of artificial intelligence on the labor market: over the past three years, wage growth in jobs highly susceptible to AI has been significantly lower than that of workers in less AI-affected industries.

The two economists calculated that from 2023, following the launch of ChatGPT in November 2022, to the present, wage growth in high-AI-risk jobs was 6.7 percentage points lower than in low-risk jobs. However, contrary to popular expectations, the degree of AI impact on jobs showed almost no correlation with job employment rates.

In their report, the economists wrote: "Various evidence suggests that the first quantifiable labor market impact of AI adoption is a narrowing of wage growth, rather than job displacement; the most economically vulnerable workers bear the brunt of this pressure."

The report listed several research limitations: the current AI job risk rating relied solely on data from Anthropic, the only frontier AI lab that publicly releases detailed usage data to support academic research. Labor data after 2023 may still retain disruptions from the COVID-19 pandemic, which could interfere with baseline statistical results.

Despite these limitations, this conclusion still provides important evidence for investors and society at large to assess the impact of the AI wave on the U.S. workforce in the coming years. The Apollo Global Management LLC team estimated that there are currently 5.8 million U.S. workers in high-AI-impact jobs, accounting for 3.7% of the total U.S. labor force of approximately 169.4 million, "far lower than estimates from various theoretical models over the past three years."

However, the economists stated: "As U.S. companies fully deepen AI implementation, the number of such jobs is likely to rise significantly, which will have profound implications for income distribution gaps and labor market policies in the future."