New Delhi: Recent college graduates from majors more exposed to artificial intelligence are finding it harder to secure jobs. They are also earning less in their first months of employment, a new US Census Bureau paper found.
According to the study, graduates from the 10% most AI-exposed majors were 5% less likely to find employment within three months of graduation than graduates from less-exposed majors. They also earned 13% less during their first three months.
The research studies how college graduates may be affected by the rise of artificial intelligence, after ChatGPT became widely available in late 2022.
How did researchers measure AI exposure?
The paper, titled Graduating into Disruption: Labor Market Outcomes for AI-Exposed College Majors, uses administrative records covering about 29% of all US bachelor’s degrees awarded between 2016 and 2024.
The study uses a 2024 study to assess tasks based on how easily they can be automated.
They then mapped those tasks to occupations and linked the occupations to college majors using data provided by universities about their students and degrees. This helped show AI exposure among the majors.
The researchers divided majors into 10 groups, or deciles, according to their AI-exposure scores. Majors in the bottom 60% of exposure were used as the study’s control group.
The exposure scores were based on the occupations graduates entered before ChatGPT became widely available in late 2022.
Importantly, the study does not suggest that highly exposed degrees or professions will disappear. It examines how graduates from those majors are performing in the labour market as AI adoption increases.
Which degrees are most exposed to AI?
Some of the majors with the highest AI-exposure scores include computer science, computer engineering, applied mathematics, statistics and journalism.
At the other end were fields such as nursing, special-needs education and medical assisting services, which had much lower AI-exposure scores.
The researchers are measuring how closely the occupations associated with these degrees overlap with AI-automated tasks. The ranking does not mean that AI will replace graduates in those fields.
Also Read: Scholars are using AI for research work, but worry about ‘AI shame’, says new study
Employment falls, but earnings take a bigger hit
The employment gap was visible soon after graduation.
Graduates from the AI-exposed majors were 5% less likely to be employed within their first three months than those from less-exposed majors.
Their earnings also fell sharply. They earned 13% less in the first three months compared to other graduates.
Where are these graduates going instead?
The decline in earnings does not appear to be simply because AI-exposed graduates are unable to find jobs.
Instead, the study found that many are moving into lower-paying industries.
“Roughly half of the decline in earnings is attributable to lower earnings within the industry sectors that employ these graduates, with the remainder resulting from a shift in the industry mix into lower-wage sectors such as restaurants and retail,” the researchers said.
Lee C. Tucker, one of the study’s authors, said graduates from the most AI-exposed majors were less likely to enter sectors traditionally associated with some of their degrees.
“Most exposed graduates are much less likely to go into the information sector, which is where a lot of the traditional computer-science jobs are, and the professional, scientific, and technical services sector,” Tucker said. He noted that these are among the highest-paying sectors for such graduates.
“Instead, they are shifting into retail, trade, and food and accommodation sectors,” he added, which are among the lower-paying sectors for these graduates.
The researchers also found that the gap narrowed as graduates moved further into their careers. But for those from the most AI-exposed majors, the effects remained substantial.
The study adds to growing research on how artificial intelligence is changing the labour market, particularly for workers trying to enter it for the first time.
Edited by Maryam Hassan