What the New Data Suggest
So far, the narrative surrounding artificial intelligence in the workplace has had one constant victim: the recent college graduate, who is supposedly the first to be laid off, as large language models automate the types of low-level tasks that traditionally provided a foundation for office careers.
New evidence, however, suggests that the technology may affect a surprisingly different segment of the workforce: those who are closer to retirement.
ΣΧΕΤΙΚΑ ΑΡΘΡΑ
Artificial Intelligence and Older Adults
In a recent study by the Center for Retirement Research at Boston College, economics professor Geoffrey Sanzenbacher carefully examined labor data, in order to map the impact of artificial intelligence on older workers in the United States.
By comparing government labor data with an AI exposure index — a dataset that tracks the extent to which certain occupations rely on tasks that AI can perform — Sanzenbacher compared the number of workers aged 55 and older who left the labor force before and after the release of ChatGPT in 2022.
Although the researcher points out that “the impact of Artificial Intelligence on any worker—let alone those nearing retirement—remains an open question,” his findings point to a concerning shift.
Since the launch of ChatGPT
Before ChatGPT, older workers in jobs most exposed to artificial intelligence—highly skilled knowledge-based jobs, such as coding and tax preparation—were more likely to continue working later in life compared to those in manual labor.
“The types of jobs exposed to artificial intelligence once had a relative advantage in terms of career longevity,” explains Sanzenbacher.
“In the post-ChatGPT era, the nearly flat trend lines for “not working” suggest that this advantage has diminished significantly, with a significant portion of the increase attributable to unemployment.”
Although this general pattern still holds true overall, Sanzenbacher notes that since the release of ChatGPT, the percentage of workers aged 55 and older leaving office jobs has been rising rapidly.
This is not necessarily because people are retiring early. As Sanzenbacher writes, “following the release of ChatGPT… jobs exposed to artificial intelligence saw relative increases in overall transitions out of the workforce and specifically into unemployment (but not out of the labor force).”
In other words, more and more older office workers are being pushed out of the plane before they’re ready to jump—with a parachute.
Who Is Affected
As with similar patterns observed in the entry-level job market, this seems to affect older office workers in particular.
For workers over the age of 55, the data suggest that the following career paths could see the most job losses.
1. Computer Programmers
Computer programmers have some of the highest “exposure to artificial intelligence,” which researchers assessed based on how well AI can perform various tasks related to their daily workflow.
Programmers were 25% more likely to become unemployed following the release of ChatGPT.
2. Accountants and Auditors
Accountants and auditors were another group facing high exposure to artificial intelligence. Researchers estimated a 22% increase in labor force exits starting in 2022.
Several careers that involve a great deal of work with numbers and the organization anddata analysis ended up on the list of those with the highest levels of exposure to artificial intelligence, including database architects and data scientists.
3. Management Analysis
Although the job relies heavily on data analysis and formulating recommendations based on these figures, management analytics roles were considered moderate in terms of exposure to artificial intelligence.
The increase in workforce turnover was estimated at approximately 16%.
4. Maintenance and Repair
Maintenance and repair jobs across all sectors saw an estimated 9% increase in workforce attrition since the rise of ChatGPT.
Although these jobs are often considered more hands-on, maintenance work today can involve machinery and other technology, which could be affected by automation tools.
These jobs were also rated as moderate in terms of exposure to artificial intelligence.
Which jobs are least affected?
While the jobs most affected, according to the researchers, involved web development and data analysis, the jobs least frequently affected involved more manual labor.
Loading and unloading workers, painters, nurses, and healthcare aides in nursing homes were ranked very low on the list of jobs at risk from artificial intelligence.
The researchers estimated that only about 1.5% of painters—a job considered to have very low exposure to AI—left the workforce.
Source: in.gr
