The Job Market is Telling Us Something About AI and Jobs…

But it’s not telling us the same things as the AI hypesters are telling us.

Last week, an AP article reported that

Young people graduating from college this spring and summer are facing one of the toughest job markets in more than a decade. The unemployment rate for degree holders ages 22 to 27 has reached its highest level in a dozen years, excluding the coronavirus pandemic. Joblessness among that group is now higher than the overall unemployment rate, and the gap is larger than it has been in more than three decades.

Hucksters for ChatGPT and other large language models (LLMs) have been quick to interpret sluggish hiring as evidence that AI is replacing human workers. We’ve heard that song-and-dance before. For example, nine years ago, Turing Award Winner and future Nobel Prize Winner Geoffrey Hinton famously said, “People should stop training radiologists now,” because it is “completely obvious” AI will outperform human radiologists within five years. Fast forward to today and the number of radiologists has actually increased.

Is it different this time?

The Internet is abuzz with personal stories of students who majored in computer science and can’t find jobs. Perhaps the AI created by computer scientists is replacing computer scientists. Amazon CEO Andy Jassey told his employees, including software engineers, that “AI will replace some of you.” An Atlantic article last week was titled, “The Computer-Science Bubble Is Bursting: Artificial intelligence is ideally suited to replacing the very type of person who built it.”

Deliciously ironic but not really true. An AMD article in the IEEE Spectrum concluded that “the kinds of tasks coding assistants are good at — namely, busting out lines of code — are actually a very small part of the software engineer’s job.”

There are two more likely explanations for the poor job prospects of computer science majors. One is that the revenue from ChatGPT and others LLM has been disappointing. AI companies can’t make profits hiring people who don’t generate revenue. Another explanation is that there are just too many CS majors. When students rush en masse to majors with the highest pay and best job prospects, they undercut the pay and job prospects they are seeking. This vicious cycle has happened with law, consulting, and finance. Now it is happening with computer science. The  Wall Street Journal reported that, “Between 2018 and 2023, the number of students majoring in computer and information science jumped from about 444,000 to 628,000 [a 40% increase].” The 2024 and 2025 numbers are surely even higher.

In addition, the bleak job market for recent college graduates extends far beyond CS and involves another set of issues that encompasses more than CS. Consider these two factoids: (1) hiring has been more robust for jobs that don’t require college degrees; and (2) many firms now prefer retaining older workers to hiring younger ones — creating a no-hire, no-fire economy.

Bias against Gen-Z job seekers

Recent college grads are Gen-Zers who had their high school education disrupted by the COVID pandemic; had their college education sabotaged by ChatGPT and other LLMs; and have attracted unflattering stereotypes.

A survey of 1,000 hiring managers found that 40% admitted to be biased against Gen Z applicants and 41% advised them not to put their graduation year on their resumes. Another survey found that “58% [of employers] say recent college graduates are unprepared for the workforce” and that 38% “prefer hiring older workers over recent college graduates.” Still another survey found that 31% of employers “avoid hiring Gen Zers and would prefer to hire older workers” and that “30% said they’ve had to fire a Gen Z worker within a month of their start date.”

Gen Zers are thought to have an entitled attitude and weak work ethic. Many are perceived to be ill-prepared for jobs that require the critical thinking and communication skills they should have acquired in college.

A college degree used to be the path for upward socioeconomic mobility. In part, college was a screening mechanism. An ability and willingness to do what was needed to obtain a college degree was tangible evidence that a person was competent, willing to work hard, and able to meet deadlines.

For the last decade, however, the brand image of colleges, particularly elite universities, has been on a downward spiral, and books about the problems exploded in the early 2010s. Fewer companies respect college degrees and thus fewer students have been attending colleges. Enrollment was down 15% between 2010 and 2021.

There is a long list of problems. College administrators are focused on retention and graduation rates — which means keeping students happy and making it easy for them to pass their courses. One outcome is the lament that “Students now view deadlines as suggestions rather than requirements.” Another is rampant grade inflation. It used to be that average students got C grades. Now, average students at private colleges get A grades. No surprise that some employers have become less impressed by college degrees and may even view a recent college degree as a negative — evidence that applicants have picked up bad habits (like entitlement and a weak work ethic) and are not prepared for jobs that require critical thinking and communication skills.

Cheating has also become more common. Students have been using answers from Chegg and other startups to do their homework for the last ten years and LLMs have enabled them to take this cheating to a whole new level in the last two years. Most term papers are written with LLMs, thus enabling students to offload their thinking to them. One UCLA graduate even bragged about his use of ChatGPT at graduation, showing how he did these term papers in a viral video.

Cynical yes, but perhaps an uncomfortable truth as well. The job market for recent college graduates may be warning us that participation trophies for attending college and the use of ChatGPT and other LLMs in lieu of critical thinking do have long-run costs.

Read More...