AI may be nudging some older workers into early retirement, study finds

AI has sparked fears about replacing the next generation of workers. New research suggests it may also be speeding up the exit of the last.

A new research brief from the Center for Retirement Research at Boston College found that workers ages 55 and older in occupations with greater exposure to AI have become more likely to leave their jobs since the launch of ChatGPT in late 2022. While the findings do not prove AI is directly causing those departures, researchers say the trend is one policymakers should watch closely.

The analysis combined data from the U.S. Current Population Survey with the Digital Planet AI Exposure Index, which measures how much of a job’s work could potentially be performed or assisted by AI. Researchers compared employment patterns before and after ChatGPT’s debut to examine whether AI-exposed occupations experienced different job exit trends.

The study found that older workers in highly AI-exposed occupations, including computer programmers and accountants, experienced larger increases in job exits after ChatGPT’s release, particularly transitions into unemployment.

Read more: Trust & Will cuts staff amid restructuring, AI disruption

Researchers say AI exposure does not necessarily mean a job is at greater risk of disappearing. Instead, the index measures how much of a job’s tasks could be completed with AI. In some cases, that can make workers more productive by automating routine work. In others, it may reduce demand for certain tasks or encourage workers to leave if they’re reluctant to adopt new technology.

Before ChatGPT’s launch, workers in highly AI-exposed occupations generally had lower job exit rates than workers in less-exposed, more physically demanding roles. Since AI adoption accelerated, however, that advantage has narrowed as exits from those occupations have increased.

Read more: AI’s uncomfortable question for advisors: Who still needs you?

The report illustrates the difference across occupations. Computer programmers, one of the jobs with the highest AI exposure, saw a predicted increase in transitions out of work more than 25% after ChatGPT’s launch. Painters, whose work is considered far less exposed to AI, saw only about a 2% increase.

Even with those increases, researchers found that many AI-exposed occupations still have lower overall job exit rates than physically demanding jobs. The report says those occupations often require more education, offer higher pay and are less physically demanding, factors that have traditionally supported longer careers. Still, the researchers said AI could narrow that long-standing advantage.

The authors cautioned that AI is still developing and other economic factors could also be affecting employment patterns, making it too early to draw firm conclusions about the technology’s long-term impact.

The findings suggest AI’s influence on the labor market extends beyond concerns about younger workers entering the workforce. As policymakers look for ways to encourage longer careers to support retirement security and Social Security, researchers say AI may be creating new challenges for some workers approaching retirement.