Contrary to common assumptions that younger workers bear the brunt of AI-driven job disruption, new findings show that older employees close to retirement are facing significant pressure in white-collar occupations. A study by the Center for Retirement Research at Boston College analyzed labor data to assess how AI exposure correlates with workforce exits among Americans aged 55 and older.
The research compared occupational data before and after the launch of ChatGPT in 2022, focusing on roles highly susceptible to AI automation—such as programming, tax preparation, and auditing. Historically, older workers in these skilled knowledge jobs tended to remain employed longer than those in manual labor positions. However, this trend shifted dramatically post-ChatGPT, with many older white-collar workers leaving their jobs not by choice but due to increased unemployment.
This phenomenon is not primarily early retirement. Instead, it reflects a rise in involuntary job losses, with displaced workers struggling to remain in the labor force. For example, workforce exits among programmers surged by more than a quarter, while accountants and auditors saw exits rise by around 22 percent since 2014. In contrast, manual labor sectors like painting experienced only modest increases in retirements during the same period.
The study highlights a widening gap in job security between older professionals in AI-exposed fields and those in less automated sectors. As hiring of entry-level workers slows and seasoned employees exit prematurely, the labor market faces a dual squeeze. This raises concerns about how workers across age groups can adapt amid accelerating technological change.