AI Accelerates Early Retirement for Workers Over 55

What Happened?

A study by the Center for Retirement Research at Boston College, led by Geoffrey Sanzenbacher, reveals that workers aged 55 and older in occupations with high exposure to artificial intelligence are leaving the labor market at a higher rate than before the launch of ChatGPT in November 2022. The research, reported by The Next Web and based on data from the U.S. Current Population Survey (CPS), found that the exit rate for these workers increased significantly in 2023 and 2024. According to the study, the probability of an older worker in high AI-exposure occupations leaving employment rose by 20% compared to the pre-ChatGPT period. Sanzenbacher notes that this phenomenon is not observed in workers under 55, suggesting a specific effect on the senior population. The research classified occupations by their exposure to AI using metrics of cognitive tasks that can be automated by tools like ChatGPT, including roles in writing, programming, data analysis, and legal services.

Why Is This Important?

This finding contradicts the prevailing narrative that AI primarily affects well-paid white-collar jobs. While it's true that higher-paying jobs are among the most exposed, the impact on older workers is particularly concerning. Unlike younger workers, who can retrain or change careers, seniors face barriers such as age discrimination, lack of updated digital skills, and less time to recoup training investments. Moreover, forced early retirement can erode retirement savings and increase pressure on social security systems. According to Federal Reserve data, 25% of Americans over 55 have no retirement savings, and an early exit from the labor market could worsen financial insecurity. The study also highlights that the most exposed occupations include not only technical jobs but also administrative and office roles, where many older workers are employed. For example, secretaries, legal assistants, and accountants face a growing risk of being replaced by AI assistants. Age discrimination, documented by AARP, shows that 61% of older workers have witnessed or experienced age discrimination at work, making it harder to find new jobs.

Historical Context and Comparisons

This is not the first time a disruptive technology has altered the labor market for older workers. Industrial automation in the 1980s and 1990s displaced manufacturing workers, many of whom were seniors. However, generative AI is different: it directly targets cognitive tasks such as writing, data analysis, and programming, often performed by experienced professionals. The speed of change is also unprecedented: while automation took decades, the adoption of ChatGPT and similar tools has been measured in months. According to a 2023 McKinsey report, generative AI adoption could accelerate the automation of up to 60% of tasks in some occupations by 2030. In comparison, robotic process automation (RPA) took nearly a decade to achieve significant adoption. Moreover, previous studies on AI's impact on employment, such as Frey and Osborne (2013), focused on manual jobs, while generative AI directly threatens knowledge workers. The case of older workers is especially relevant because, unlike younger workers, they have less flexibility to adapt. A 2021 OECD study found that workers over 55 are 40% less likely to participate in training programs than younger workers, exacerbating their vulnerability.

Consequences for Companies and Workers

For companies, the premature exit of older workers means a loss of institutional knowledge and experience. On the other hand, it may reduce short-term salary costs, but in the long run, it could lead to talent shortages in roles requiring judgment and experience. A Deloitte report estimates that the loss of tacit knowledge from older workers could cost U.S. companies up to $500 billion annually in lost productivity. For workers, the news is alarming: those planning to work until age 65 or 70 may be forced to retire earlier, with less savings and benefits. According to the Center for Retirement Research, the early retirement rate among workers aged 55-64 in AI-exposed occupations rose from 15% in 2022 to 22% in 2024. This could increase reliance on Social Security, which already faces funding deficits. Policymakers should consider retraining programs specifically for those over 55 and strengthen social safety nets. Countries like Singapore have implemented training subsidies for older workers, while the European Union has launched digital upskilling initiatives. However, in the U.S., existing programs like the Workforce Innovation and Opportunity Act (WIOA) are not designed to address the needs of workers displaced by AI.

What Should Readers Know?

• AI exposure doesn't only affect graduates; senior workers are particularly vulnerable, even in non-technical white-collar occupations.
• The early retirement rate in AI-exposed occupations has increased since 2023, with a 20% rise in the probability of exit.
• Age discrimination and lack of retraining opportunities worsen the problem; only 10% of older workers receive employer-sponsored training.
• Companies and governments must act to mitigate the social and economic impact, for example, through tax incentives for hiring older workers or reverse mentoring programs.
• The study underscores that AI is not neutral: its effects are unevenly distributed, and older adults are a high-risk group requiring urgent policy attention.

“Older workers are being pushed out of the labor market by AI, and this is just the beginning,” warns Sanzenbacher. “We need policies that recognize their value and provide support for the transition.”

The study highlights the need for inclusive policies that address the specific needs of older workers in the AI era. Without intervention, the skills gap and economic inequality could deepen, affecting both individuals and the economy as a whole.