Why Are Male Non-STEM Graduates Struggling to Find Jobs?—In the AI Era, What You Did at University Matters

Recently, I came across a discussion about how university students in non-STEM fields, particularly men, are finding the job market increasingly difficult.

STEM students, meanwhile, are apparently benefiting from an unprecedented candidate-driven job market in which employers compete for graduates.

One explanation given was the reduction of administrative work through generative AI and IT. Work previously assigned to junior employees—preparing documents, performing routine administrative tasks, and assisting sales teams—is declining, narrowing the entry point into the generalist office roles traditionally associated with non-STEM graduates in Japan.

Another analysis suggested that the remaining roles place greater emphasis on communication skills and that female candidates are more likely to be selected.

But as someone with a STEM background, I find it strangely unsurprising that young non-STEM graduates are struggling.

Rather than being a problem that suddenly appeared with AI, perhaps AI has merely made the long-standing relationship between university education and employment impossible to ignore.

People Have Always Said That STEM Graduates Have an Advantage in Employment

Even when I was a student, people often said:

“If you want a job, studying a STEM subject gives you an advantage.”

At the time, my reaction was roughly:

“Well, of course it does.”

We conducted experiments, wrote reports, went to the laboratory, and then conducted more experiments.

Meanwhile, my friends in non-STEM fields joined student clubs, worked part-time jobs, traveled, and appeared to be enjoying university life.

Then, at graduation, all of us entered the job market as holders of the same bachelor’s degree. From the STEM side, I felt we deserved at least some reward for the difference.

That said, STEM graduates do not necessarily have an employment advantage because the specialized knowledge they acquired at university can be applied directly in a company.

The more important reason lies elsewhere.

The Undergraduate Thesis Itself Is Not Particularly Valuable

To put it bluntly, an undergraduate research project completed in a student’s fourth year is little more than play from the perspective of professional researchers.

A master’s thesis is somewhat more serious, but companies still do not hire a student because they think:

“This is outstanding research. We would love to use it at our company.”

What has value is not the finished paper but what the student did to reach it.

Research begins by formulating a question. You examine previous research, develop a hypothesis, conduct experiments or calculations, and look at the results.

If the results differ from your prediction, you consider why.

Even when the data is inconvenient, you cannot pretend it does not exist.

Finally, you organize the work into a paper that you can defend when others challenge it.

In other words, a university laboratory repeatedly trains students not only in specialized knowledge but also in how to:

Formulate their own questions, gather facts, test them, and reach a coherent conclusion.

I found this training far more useful after entering the workforce.

When “I Spent Four Years at University Having Fun” Was Still Acceptable

Of course, some non-STEM students study intensely, and some STEM students barely study at all.

Still, when I was a student, it was perfectly common for my friends in non-STEM fields to struggle to explain what they had studied at university when someone asked them.

They spent four years with considerable freedom, wrote something good enough to graduate at the end, and entered the workforce.

In the past, that was not a problem.

Companies were hiring large numbers of young people.

Even without specialized skills, graduates had entry points into sales, administration, and similar roles. Companies could hire them straight out of university and teach them the work after they joined.

In a sense, companies took responsibility for continuing their university education.

Then AI began moving into that entry point.

AI May Be Eliminating the Tasks That Came Before the “Professional” Work

Summarizing documents. Taking meeting minutes. Searching for information. Writing routine text. Answering simple inquiries.

Generative AI is extremely good at the kinds of tasks that were easy to assign to junior employees.

From an employer’s perspective, this makes the cost of saying, “Let us hire this young person even though they cannot do anything yet, give them simple work, and spend five years developing them,” feel heavier than it once did.

The irony is that although AI has made the task of searching for information easier, the ability to judge whether the information found is correct has become even more important.

AI can generate an endless supply of plausible answers.

That is precisely why we need people who can ask:

“Is that assumption true?”

“Where is the primary source?”

“Can that conclusion really be drawn from this data?”

And this is exactly the kind of work students are forced to do over and over again in research.

In the AI Era, the Chair Itself Is Disappearing

In the past, even students who spent four years at university without studying very much still had a fairly common route into a company that would train them after they joined.

AI is beginning to remove that grace period.

We often hear that “AI will take people’s jobs.”

For new graduates, however, the situation may be slightly different.

They have not yet been given a job, so it has not even been taken away from them.

A student spends four years at university doing little but having fun, begins job hunting with the expectation that “the company will train me once I start working,” and discovers that AI is already doing the beginner-level work.

The job was not taken away. The chair was removed before the person even had a chance to sit down.