AI might make the liberal arts matter more

The more I listen to people talk seriously about AI, the more I keep coming back to education.

Dr. Alexander Wissner-Gross and Dave Blundin are two people I follow closely. A lot of their conversations touch on what happens when AI can do more of the work we once assumed required years of specialized training.

Blundin has argued that AI is becoming a foundation for almost every field, from biotechnology to space. He has also talked about how quickly college curriculums can fall behind when the technology changes faster than a university can update what it teaches. Wissner-Gross is interesting to hear on this because he is certainly not coming from an anti-education position. He earned three undergraduate degrees at MIT, then a physics Ph.D. at Harvard.

Listening to people like them has made me wonder what a degree is really giving us if AI keeps getting better at the specialized work those degrees were meant to prepare us for.

It is not just about the subject

That question came up recently in conversations with friends, and we kept circling back to liberal arts. Not because we think an English degree or a history degree suddenly guarantees a good career. It does not.

What stood out was that a liberal arts education is supposed to do more than teach a subject. At its best, it teaches you how to read carefully, question assumptions, research something unfamiliar, explain your thinking, understand another point of view, and work through ideas that do not have simple answers.

Most of that can be summed up as learning how to learn.

The American Association of University Professors puts some shape around that idea. It describes critical thinking as a collection of overlapping skills: recognizing assumptions, drawing inferences, judging the authority of an argument or source, and dealing honestly with complexity. Those are useful skills in any field. They are especially useful when a tool can hand you a polished answer in seconds.

The statistic was not quite right

Then I heard a statistic that pushed this from a conversation into something I wanted to write about. I remembered it as something close to, “half of Silicon Valley CEOs have liberal arts degrees.”

That is not what the research says.

A Kauffman Foundation study looked at 652 U.S.-born CEOs and heads of product development from 502 engineering and technology companies founded between 1995 and 2005. It found that 37 percent had terminal degrees in engineering or computer technology and another 2 percent had mathematics degrees. The rest were spread across business, applied sciences, health care, law, economics, arts and humanities, and other fields.

That is a much less dramatic claim, but I think it is the more useful one. The technical leaders in that sample did not all come through one narrow educational pipeline. The study is older, and it cannot tell us which degree made anyone a better leader. Still, it is a useful reminder that technology has always needed more than people who can do one technical thing well.

A separate study of the undergraduate majors of the 2019 Fortune 50 CEOs found an even split: 18 STEM majors, 18 business majors, and 18 liberal arts majors. One CEO attended college but did not graduate. That does not prove liberal arts graduates make better CEOs. It does make the usual story about one correct major look a little thin.

Most of the work is not code

That feels familiar to me after spending most of my career in technology. Technical knowledge matters, obviously. But a surprising amount of the work is not really about code.

Someone has to understand what the user is actually asking for. Someone has to notice when the requirement itself is wrong. Someone has to explain an idea clearly enough that designers, developers, managers, and users all understand the same thing. Someone has to recognize when a project is solving the wrong problem.

I have worked on plenty of projects where writing the code was easier than getting everyone to agree on what we were trying to build.

AI can make an answer easier to get. It cannot make a good question appear on its own.

AI changes the value of the question

This is where it connects back to AI for me. It is already lowering some of the barriers between having an idea and building something from it. I can use it to write code, explain technology I do not know well, analyze information, research a subject, create an interface, or help me work through a problem outside my normal area of expertise.

That does not make expertise unimportant. AI can be wrong, and it can be confidently wrong. Knowing enough to recognize a bad answer still matters. So does knowing when an answer is incomplete, when a source does not support the claim, or when the whole framing of a question needs to change.

But it may change which skills become more valuable. If access to specialized information becomes easier, then knowing how to ask a good question starts to matter more. So does knowing how to judge an answer, connect ideas, explain what you mean, and keep learning when the tools change.

Those are not exclusively liberal arts skills. They should not be. Engineers, scientists, designers, tradespeople, and everyone else need them too. But they are a big part of what liberal arts education is supposed to develop.

The old distinction may matter less

For years, liberal arts degrees have often been treated as the less practical choice, while STEM was presented as the safer path into a good career. AI may not reverse that. I am not sure anyone knows yet.

But it might make the old distinction less meaningful. The question may become less about what someone studied twenty years ago and more about whether they still know how to learn something new today.

That feels like a more durable kind of education to me.