Via apollo.com
High-AI-exposure occupations saw 6.7% slower wage growth while employment levels barely budged, according to new research from Apollo's chief economist
The robots aren’t taking your job. They’re just making it pay less.
A new study from Apollo Global Management’s Chief Economist Torsten Sløk and analyst Sania Edlich found that AI’s earliest measurable impact on the labor market is showing up in paychecks, not pink slips. Across 321 matched occupations tracked from 2015 to 2025, jobs with high AI exposure saw real wage growth come in 6.7% below their low-exposure counterparts after 2023. Employment levels, meanwhile, showed no statistically significant change.
Who’s getting squeezed #
The wage hit isn’t landing evenly. Service workers faced a relative wage decline of 24.3%, the sharpest among occupational categories studied. Workers in the bottom earnings quartile saw wages drop 10.7% relative to peers in low-exposure jobs.
Top earners? No significant effect.
The study used a difference-in-differences methodology, a standard econometric approach that compares changes in outcomes between a treatment group (high AI exposure) and a control group (low exposure) before and after a specific event. In this case, that event was the explosion of generative AI capabilities beginning in late 2022.
Crucially, the researchers didn’t rely on theoretical guesses about which jobs AI might affect. They used real usage data from the Anthropic Economic Index, which tracks actual AI deployment across roughly 800 occupations matched to Bureau of Labor Statistics codes.
The $28 billion question #
Apollo’s team conservatively estimated the aggregate annual labor-income impact at approximately $28 billion, affecting around 5.8 million workers. Spread across 5.8 million workers, it averages out to nearly $4,800 per person per year in lost wage growth.
The mechanism appears to be supply-side pressure. When AI tools can perform portions of a job, employers have more leverage in wage negotiations. They don’t need to fire anyone to capture the savings. They just need to hold the line on raises, knowing the alternative to human labor is getting cheaper every quarter.
The other side of the ledger #
Sløk’s research also highlighted that AI is helping fuel a surge in new business formations. The logic is straightforward: AI lowers startup costs. Tasks that once required hiring specialists, from building a website to drafting legal documents to running basic market analysis, can now be handled by a founder with the right prompts and a subscription. That reduction in barriers to entry is showing up in the data as record levels of new business applications.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our