MW Choosing these AI-exposed college majors could dent your job prospects - and lower your pay

Andrew Keshner

'AI is becoming a lens they use to judge whether a major is a safe bet.'

At a time when many college students say they are rethinking their academic focus because of AI, new findings supply hard data on the potential costs of these decisions.

Picking a major has long been a weighty decision for college students, but the financial stakes appear to be getting higher in the AI era, according to economists at the Federal Reserve Bank of Dallas.

At a time when many college students say they are rethinking their academic focus because of the fast-emerging technology, new findings supply hard data on the potential costs of these decisions.

Students whose majors have traditionally been a pipeline to jobs that can easily be automated were less likely to find a job within a year of graduating, Dallas Fed economists wrote Tuesday. When these graduates landed work, they were also likely getting paid less compared to peers who had majors that were less vulnerable to AI's advance in the workplace.

Researchers dug into administrative and earnings records of recent graduates from public Texas universities. The divide became clear after 2022, the year when OpenAI launched ChatGPT.

Students with more AI-exposed majors had a 1.7-percentage-point drop in their chance of finding work in the state within a year, compared to those with less AI-exposed majors. Put differently, researchers said May 2024 graduates with more AI-exposed majors were roughly 8% less likely to find a job within the first year. These students were also more likely to end up seeking a graduate degree, they noted.

Once finding work, graduates with more AI-exposed degrees earned 5% lower wages compared to classmates with less AI-exposed majors, the findings showed.

Supposing a student had four quarters of work straight out of school, May 2024 graduates stood to make $3,000 less for every 10% increase in how much of their major's skills could be open to automation in a work setting.

There are different ways academic research defines AI "exposure." Dallas Fed researchers made their measures with a database of jobs and their associated tasks, and records of the tasks that Anthropic's Claude can perform.

Some of the most AI-exposed majors included computer science, computer engineering and language majors, researchers said. Other highly exposed majors included management information systems and majors related to rhetoric, composition and writing.

Meanwhile, nursing, education and psychology scored as some of the least exposed to AI encroachment. Social work, animal sciences as well as hospitality and administrative management also faced little exposure.

Indeed, computer-science and engineering majors had 2024 jobless rates ranging between 7% and nearly 8%, according to the New York Fed. Nursing and education majors had jobless rates between 1% and approximately 2%. Special-education majors had a 0.7% jobless rate, which was the lowest in the class of 2024.

Other researchers are also tracing the link between college majors and job-market outcomes.

Students graduating with majors that were most exposed to AI in the workplace had a 5-percentage-point drop in their chance of employment in the quarter after graduation, according to a U.S. Census Bureau working paper released this month. They experienced an initial full-quarter earnings decline of 13% that would be comparable in scale "with graduating into a large recession," the authors wrote.

AI's ascent has been a puzzle for students, college professors and administrators who are trying to figure out if they should be embracing the technology or avoiding it.

Career centers for nearly 60% of colleges and universities say they have started communicating with students about how AI may impact careers, and not just in the tech sector, according to a spring survey from National Association of Colleges and Employers.

Conversations about AI's impact are happening before students step on campus, said Stephanie Simpson, CEO of the Independent Education Consultants Association, whose members work with families on college decisions.

"More students and parents are asking whether the careers they're pursuing will still exist in five or 10 years, and how AI impacts the return on their investment in the college experience. For many, AI is becoming a lens they use to judge whether a major is a safe bet," said Simpson.

Schools are coming up with a flurry of AI majors, minors, concentrations and certifications, which raises questions about what even counts as a strong AI program these days, Simpson noted.

"Right now, it's not entirely clear what the quality benchmark is," she said. "The field is changing so fast that the standard feels like a moving target."

Read also: Job seekers can now spend thousands on AI certificates. Is it worth it?

Students, parents and higher-education experts are grappling with AI's impact on campus. Yet breaking into the labor market can be a tough job, regardless of whether your skills can be replaced by AI.

Last month, there were 245,000 people between the ages of 20 and 24 with no prior work experience who were unemployed, according to the Bureau of Labor Statistics. That's the largest August number for that demographic since 2014 - and it's often a month when new grads are starting their first job, one economist previously told MarketWatch, also pointing to AI's impact.

Despite rising anxiety about AI, the technology hasn't yet resulted in massive job losses. Instead, the job market is largely stuck in a low-hire, low-fire speed. That helps current employees who may be holding on to their jobs, yet it's rough for newcomers who are on the outside looking in.

It's now been 35 straight months where employment has been shrinking for 22- to 25-year olds in jobs with high exposure to AI, according to new numbers Wednesday from ADP (ADP) and Stanford University researchers. That's on a year-over-year basis. The trend started in October 2023, one year after ChatGPT's arrival, they noted.

There's no guarantee that so-called AI-proof jobs are a silver-bullet solution for younger workers. In August, ADP and Stanford researchers found a 2% year-over-year drop in employment for early- to mid-20s workers in the least AI-exposed jobs.

For all age groups, employment in jobs with high AI exposure fell 0.6% from a year ago, and employment in the least-exposed jobs increased 0.2%, researchers said.