AI may be shortening some older workers’ careers, research finds

AI is starting to change the retirement math for some older workers. New research from the Center for Retirement Research at Boston College finds that workers age 55 and up in jobs more exposed to artificial intelligence are leaving work more often than they did before ChatGPT arrived.

That matters for households and policymakers because working longer is one of the main ways people can boost retirement security. If AI makes some higher-earning careers harder to extend, it could complicate future changes to Social Security, according to Geoffrey Sanzenbacher, the Boston College economics professor who wrote the paper.

How AI could affect older workers

The paper defines AI exposure as the degree to which artificial intelligence can perform tasks in a given occupation. It uses Current Population Survey data and AI exposure scores from Tufts University’s Digital Planet initiative, which studies digital technology and its economic effects.

Sanzenbacher told CNBC that workers 55 and older in AI-exposed industries are moving out of jobs more often, with the exits split between unemployment and voluntary departures. He said the effect is statistically significant, meaning the pattern is unlikely to be random, and can be sizable in some occupations.

The research lays out three possible paths. AI could automate tasks and push some older workers into unemployment or out of the labor force. The need to learn new AI tools could also lead some workers to switch into roles with less technology pressure or retire. A third possibility goes the other direction: generative AI, meaning tools that create text, code, images or other output, could make some jobs more productive and help people keep working longer.

Before OpenAI released ChatGPT, older workers in AI-exposed roles were less likely to leave their jobs, according to the research. After ChatGPT’s launch, they became somewhat more likely to leave work, including through unemployment.

Which jobs are most exposed

The jobs with the highest AI exposure in the Digital Planet data are concentrated in technology and data work. According to Sanzenbacher’s research, the five most exposed occupations are:

• Web and digital interface designers
• Web developers
• Database architects
• Computer programmers
• Data scientists

The jobs with the lowest AI exposure are more physical or hands-on. The research lists these as:

• Excavating and loading operations and mining workers
• Roof bolters and mining workers
• Orderlies
• Painting and spraying workers
• Fiberglass laminators and fabricators

Sanzenbacher found that older workers in more AI-exposed roles are more likely to be white, college-educated and higher earners than workers in low-exposure jobs. That challenges the usual assumption that physically demanding jobs are the main reason some workers cannot stay employed until later retirement ages.

The Social Security connection

The timing is sensitive for Social Security. The program’s trustees project that the trust fund used to help pay retirement benefits could be depleted in late 2032.

Lawmakers have several options to address the shortfall, including raising the retirement age or increasing payroll taxes on higher earners. The retirement age was previously raised from 65 to 67 under reforms passed in 1983.

Sanzenbacher told CNBC that higher-income people have a strong chance of facing larger benefit cuts than lower-income people under future Social Security changes. Those are also the workers who may need to keep working longer, he said, even as AI changes the jobs many of them hold.

How workers are responding

AARP research found mixed feelings among older adults. In a March survey of 1,015 adults age 50 and over, 24% said AI was a threat to their line of work, 19% called it an opportunity and 37% said it was both.

Separate research from AARP and LinkedIn found experienced professionals were more likely than younger workers to hold jobs insulated from generative AI disruption, at 49.4% versus 42.2%. The research said older workers are more likely to use skills AI has trouble replacing, including collaboration, judgment and leadership.

Vicki Salemi, a career expert at Monster, told CNBC that older professionals who have not started using AI still have time. Monster’s December WorkWatch report found 42% of 1,504 surveyed workers do not use AI at all. Among those who do, common uses include email, scheduling, writing support, coding, automation, data analysis, job applications and creative work.

Salemi said workers can start with AI tools their employer already uses while continuing to show interpersonal workplace abilities such as communication, relationship-building and problem-solving.

This story draws on original reporting from CNBC.