Why I Am Against STEM Education*

I believe that today’s emphasis on STEM (Science, Technology, Engineering, and Mathematics) education is wrong, for it ignores other critical topics, it emphasizes the teaching of specializations, it ignores the need for cross-disciplinary collaboration, and it emphasizes teaching, not learning. I speak against STEM, even though I am highly educated in S, T, E, and M. Moreover, the problems with STEM education are true of all of current education.

What are the problems with STEM education? Four major factors:

1. The emphasis on “teaching” rather than learning

2. Emphasizing STEM but overlooking the relevance of the resulting understanding to human life and society.

3. The focus on individual areas of specialization

4. The insistence on individual rather than collaborative work.

The modern understanding of learning, or learning theory, emphasizes the importance of active learning, where students engage in working through issues for themselves, guided by their instructors. Lectures are excellent vehicles for presenting overviews of topics, motivating students, and illustrating the connections between different topics. Lectures, however, are especially weak in delivering an in-depth understanding of topics and skills because students are passive, in theory absorbing the material but without the need to apply it in depth. To learn, students must be active learners, solving problems and applying the knowledge they have acquired. This often means that they will face difficulties, struggle with some areas, and make errors in others. These struggles strengthen understanding. Lectures, of course, are usually accompanied by reading assignments and homework, and it is during the homework phase that true learning can take place. However, without continual interaction with instructors, student difficulties and errors, which could be opportunities to deepen understanding, are not given sufficient attention. The use of AI tutors is controversial, but I believe that they will be very helpful, if only because human tutors are not available 24 hours/day, every day of the year: see (Khan, 2024)

A good example comes from the work of Seymour Papert, an MIT Professor, who invented a simple computer language, Logo, (Papert, 1993)) that could be used to control a simple robotic “turtle” to move across a piece of paper (or a computer screen) so that its path created a picture (the turtle had a pen attached to its tail, thus marking its path – the turtle would be placed on top of a large piece of paper. The system was designed to be easy to learn, but the problems assigned to students were deliberately designed to induce errors, even for senior MIT professors. Why would he deliberately lead students into error? Because Papert insisted that the most important thing to learn was to use an error as a tool. He called this “debugging,” after the process that computer programmers use to understand why their program doesn’t work the way they expected it to.

Here is one of Papert’s favorite examples, one that he deliberately wants people to fail.

Task: Command the turtle to draw an equilateral triangle

The answer most people give is almost always to use two commands, Forward and right. Forward tell the turtle how far to move in a forward direction. Right tells the turtle how many degrees to move to the right. (They can also use the command “Back” and “Left” but for this example I use only Forward and Right.)

The result was a sequence of instructions to the Turtle, something like

1. Forward 100

2. Right 60

3. Forward 100

4. Right 60

5. Forward 100.

(Try it yourself. Imagine following the instructions: Do you get a triangle? If not, why not?)

The Turtle would not draw a triangle: it would draw the start of a hexagon. The point of this exercise is that students should not feel ashamed of getting it wrong: instead, they should think of it as an interesting challenge – why is it wrong? Papert’s goal was not to teach how to make a triangle but rather to “debug” their thinking process so that the students would learn how to think through the problems they encounter.

Not only is this an excellent learning exercise because learning by debugging a misconception stays in the mind, but it also teaches students not to feel inferior when they make errors: experts err frequently, but it does not stop them from doing their work. The point is that we all learn best through hands-on experience. That’s one reason homework is essential. The problem with homework, however, is that it is not motivating. It is often done only because it is required, not because the students themselves have intrinsic motivation.

“to believe that only STEM drives progress and economic growth reflects something wrong in education and our public culture, a remarkable degree of ignorance about how the world really works.”Geoff Mulgan (2026).

The emphasis on STEM instruction leaves students with an excellent grasp of technical understanding but little or no understanding of how this relates to real life. Indeed, schools are downplaying the non-STEM areas of their curriculum, minimizing courses in the humanities, arts, and social sciences. But isn’t the point of STEM to help invent and devise technology for the betterment of human life? And if so, isn’t it essential to have studied these topics? Yes, quite often, a complete curriculum requires a few courses in these non-STEM fields, but these courses suffer because they are not integrated into STEM learning. Instead, the focus is on individual areas of specialization.

Traditional instruction is divided into courses, each course covering a single topic. Although this sounds logical, it can lead to weak motivation. If the topic is already interesting to the student, this approach can be effective. However, many topics are not of immediate interest, and as a result, the students do not understand why they need to learn them.

Mathematics is typically taught in stages, starting with students progressing through basic arithmetic operations to algebra, geometry, and trigonometry, and then to advanced topics such as algebra and calculus. Although instructors often try to provide examples of where the knowledge is useful, as the topics progress up the ladder of sophistication, the work seems increasingly removed from its practical utility. Again, for some students, this is perfectly fine, but others struggle to understand its relevance and, as a result, struggle to learn effectively. I recall my difficulties with calculus until the science and engineering courses I was taking began to apply the mathematics I had been taught. Then, I understood why I needed to learn it, and I appreciated the power the knowledge gave me in solving problems. So why were the topics taught separately? Why not combine the skills into one cohesive package?

In the world outside of school and academia, work is done by groups. Learning to work with others in a collaborative rather than competitive atmosphere is a critical skill in the real world but not in academia. Why are students punished for working with others, punished if they copy material done by one student into an assignment of their own, and punished if, during an examination, a student asks another student for help?

To most readers, these may seem like trivial questions. Of course, we must assess each student’s abilities so they can perform without assistance. True, but what I tell my students is that quite often, a stupid question is the most powerful question, for it questions the most fundamental assumptions in the field. And when a fundamental assumption is questioned, perhaps one can develop an alternative that has a positive impact on an entire field.1

It is obvious to everyone that the sun travels across the sky, revolving around the earth. Similarly, look at the earth: it is flat. Recognizing that the earth is not only round but also spherical so that people on the opposite side of the sphere are upside down seems incredibly stupid. Similarly, the sun appears to move across the sky, but we perceive it as moving around a stationary Earth. Although we are aware of the truth today, the assumption that the Earth is stationary remains deeply ingrained in our subconscious.

Edward Hutchins is a cognitive anthropologist with whom I used to work: see his highly influential book, Cognition in the Wild, (Hutchins, 1995). He illustrated the difficulty of changing one’s point of view by asking people to imagine it is noon and then to point at the sun. People would point straight up into the air. Then, he would ask them to imagine it is sunset and to point at the sun. They would point their hands horizontally toward the west.

Then, Hutchins would say, ‘ You pointed up and then horizontally.’ Where did those points intersect? The usual answer was puzzlement. They did not intersect. It was difficult to realize that they both intersected at a single point: the stationary sun. It was the person who had moved and rotated so that, at noon, the sun was overhead, and at sunset, the same location was now at a different angle.

Changing one’s basic assumptions about the world is difficult, but it can be done, and once done, it leads to entirely new solutions to many different issues.

Working with others, especially on homework assignments or examinations, is considered cheating. That’s why using AI systems to write essays is considered cheating. But why?

When we work in the real world, we are supposed to ask others for help. If we are writing and find that someone else has already explained the topic well, why not use their words? The reason this is considered cheating in school is that it is not allowed; therefore, students must lie and pretend that their work is their own. It is supposedly encouraged to cheat, which includes copying the work of others (or AI systems) or asking for help from others. Still, it is insisted that when this is done, credit must be given to the source (just as I have done by crediting Papert and Hutchins).

I am proud that my books build upon the work of others. Why am I proud? This demonstrates that the ideas are not simply those of a random writer (me) but are held by others. What did I contribute? First, I was selective in the material to be covered and then in providing a novel framework for the material. My contribution was the framework and perspective that I applied to the work done by a wider variety of people. It is the framework that is unique: the facts and opinions are sourced from others, with full credit given.

If this is rewarded in the world, why aren’t we teaching it in our school system? Is encouragement of cheating a stupid answer to a stupid question? If full credit is given to the source, it would not be cheating (Norman, 2005)

I just pointed out the flaws in today’s education, but that is the easy part: what is my alternative? I recommend combining disciplines, utilizing joint teaching, and requiring students to collaborate in groups. One simple way to achieve this is to teach through projects because any real project requires knowledge of multiple specialized disciplines. Moreover, because no individual can possess all the necessary disciplines, collaboration with people possessing different knowledge and skills is required.

Moreover, projects can be motivating. Working on a real issue of value to many is a powerful motivator. If the projects are real, not artificial, the motivation is further strengthened: people are awaiting the results.

Today, many fields offer each project course, but invariably, the problems are artificial, created specifically for the course. Although students often divide into groups to work on the projects, they are all from the same class. Real projects require people from other disciplines. I have witnessed (and taught) projects where all the students were from the same class, as well as projects that brought together students from several different engineering fields, business, and even the humanities. The latter is much more difficult to get going: the students all speak differently about the same topic, and quite often, there is an artificial hierarchy: business students think they are in charge, engineering students think they have all the requisite knowledge, and they from the humanities and social sciences wonder why they are there.

As the projects continue, students begin to recognize that each area of study has its valuable lessons and contributes to the final product. There is no such thing as “the most important discipline.” All are required in unison.

Yes, we still need specialties. Specialized knowledge is essential, but even the most specialized knowledge cannot exist in isolation: to achieve anything of value in the world requires an understanding of all disciplines.

Consider modern medicine. When the COVID-19 pandemic struck, experts in infectious diseases rushed to give their advice. In the absence of an effective vaccine, the major way to prevent the spread of contagious diseases is through the isolation of individuals. As a result, the scientific opinion of medical experts was to prevent people from contaminating one another, ideally through isolation.

Yes, listen to the experts, but ensure you have the advice of all relevant experts, including sociologists, anthropologists, economists, educators, and businesspeople. The toll on human life is high if people get infected, sick, and perhaps suffer permanent damage or even death. However, the toll is also high if people can no longer earn a living, if essential jobs cannot be performed, or if education is suddenly halted for an extended period at an early age.

What is the correct response? It is likely to vary across different communities, populations, occupations, and activities. Ultimately, this is a political decision made not by the opinions of politicians alone but by politicians who consider all expert opinions and determine how to combine them in a manner that is most beneficial to their community.

We must all learn to work together. No single discipline is most valuable.

Hutchins, E. (1995). Cognition in the wild. Cambridge, Mass.: MIT Press.

Mulgan, G. (2026). The myth of STEM only growth holds back the uk. LSE Blogs. https://blogs.lse.ac.uk/impactofsocialsciences/2026/04/14/the-myth-of-stem-only-growth-holds-back-the-uk/

Norman, D. A. (2005). In defense of cheating. Ubiquity, 2005 (April), 1. https://doi.org/10.1145/1066340.1066347, https://jnd.org/in-defense-of-cheating/.

Papert, S. (1993). Mindstorms: Children, computers, and powerful ideas (2nd). New York: Basic Books.