Future-Proofing Education
Summary: Traditional education is obsolete in the age of AI. Future success hinges on mastering AI tools, understanding human psychology and persuasion, and prioritizing real-world project execution over textbook knowledge.
(See also my music video about the future of skills: YouTube, 3 min. song).
AI is the cause of much wailing and pessimism in the decel and doomer communities, but most of these worries can be dismissed easily. No, AI won’t kill us all with no warning. No, AI won’t cause mass unemployment. (This latter worry is the fixed-work fallacy, which has been disproven again and again for 10,000 years of human history. When something has been true for 10,000 years, chances are that it’ll extend for another 1% of history — the next 100 years — as well.)
However, one worry is shared by many sensible people, including accelerationists like myself. We do believe that the coming decades of the AI revolution will bring untold benefits to humanity, especially in poor countries, through the widespread availability of education, healthcare, and mental health. But the more we believe that AI will change everything — and especially change business — the harder it becomes to advise young people on how to prepare for this future.
When everything is dramatically different, the old rules get defenestrated. Studying medicine to become a rich doctor with a lifelong cushy career? Out the window what that old chestnut of advice, when an AI doctor can diagnose patients with superior accuracy and a robot surgeon can fix the problem with unerring precision.
Studying medicine for a safe and prosperous future career? Out the hospital window with that traditional thinking. Medical AI and robots may be providing most healthcare in 20 years, just when today’s medical students hope to be at the peak of their careers. (Ideogram)
How can we predict which careers will be safe from AI 20 years from now, let alone in 40 years when most of today’s students will be near the end of their work life? The only answer is that we can’t. This goes double for advising parents on how to secure the best future for their 10-year-olds, given that they will end their careers in about 60 years when AI will be unimaginably better than today.
Just because something is impossible has never prevented me from trying. So I’ll give my best advice anyway, knowing that the specifics are bound to be wrong.
Three Foundational Skills
Allie Miller, who’s a leading analyst on how to use AI for business, recently posted a list of advice on what children should learn today to future-proof themselves against being made redundant by AI. She quoted 7 luminaries (and herself) from Richard Branson (Virgin Atlantic) over Sam Altman (OpenAI) to Jensen Huang (NVIDIA). In total, 24 skills were named “skills that AI cannot easily automate away” and thus worth particular emphasis in fundamental education.
The only skills listed more than once were:
Creativity (named by 4 luminaries)
AI literacy, tool familiarity, and prompting (named by 4 luminaries)
Adaptability (named by 2 luminaries)
Critical thinking (named by 2 luminaries)
Lifelong learning or high rate of learning (named by 2 luminaries)
Resilience or resourcefulness (named by 2 luminaries)
All six of these skills are surely useful and worth cultivating throughout a person’s exposure to the education system, from early education to graduate school. That said, creativity and critical thinking are absolutely not skills that “AI won’t automate away.” It’s already the case that current, primitive, AI is more creative than 99% of humans. Good luck being in that top 1% that’s currently more creative than AI: you won’t be crowing about your superiority after another two generations of AI products — about 3 years from now. The same goes for empathy, which was on the list once as a “unique human capability.” Not so, AI can exhibit empathy, and many patients prefer the empathy they’re shown by an AI healthcare provider to what they get from their human doctor.
On the other hand, adaptability, lifelong learning, and resilience/resourcefulness will absolutely be useful. The only constant is change. Furthermore, change will happen at an accelerating pace, as AI fuels innovation and changes in business and society.
Move fast, but don’t break things. Accelerating change is the only certainty about our AI-fueled future, so cultivating adaptability, lifelong learning, and resilience will be key. (Midjourney)
1: Use AI
The first skill I would teach all kids is how to use AI. By this, I don’t mean stupid prompt engineering tricks like “I’ll tip you $20 if you do a thorough job” or “this task is very important to my career.” It may currently be the case that adding such nonsense to a prompt improves output by a smidgeon from some AI tools. But surely, this will change with future releases.
In fact, it’s quite likely that future AI will be less based on long text prompts and more accessed through a hybrid user interface. A text-only UI is so 1970s.
More strategic prompting concepts, such as the prompting diamond that alternates expansion and refinement, are likely to retain their relevance much longer. The same goes for the proper use of iteration and the use of tools like inpainting in image generation. Again, we would hope for better UI control over inpainting and for vastly improved usability of these features, relative to that provided by, for example, Midjourney today. But the concepts of how to selectively control the improvement of AI-generated results will be a more important skill.
A kid in elementary school today should be taught image generation tools in art classes. Not with the expectation that he or she would use any of those same commands in a real job ten years later. But you need to start young with an appreciation of the concepts, and then the student can grow and adapt his or her skill level as the tools change, because it will only be necessary to learn each change as it happens, rather than starting from scratch with an understanding of what AI can do and how to control it.
Right now, human designers add value to AI-generated images by selecting the better-looking of the presented alternatives. But I expect that AI will add a strong sense of aesthetics shortly, including the ability to design a series of UI screens that exhibit the UX equivalent of character coherence. (Midjourney’s style references are a primitive stab in this direction.) Generative UI will soon be the way to make not just individual screens but complete workflows. Anything left for human designers? Hard to say, but the UX strategy and architecture will likely still benefit from a heavy human dose, since the goal of UX is to design for humans.
Most fundamentally, students of all ages should learn how to use AI in synergy with their own human skills. In this symbiosis, sometimes the AI serves as the originator, and sometimes it’s the refiner, and understanding how to move between these two different uses of the tool is a skill that should be grown and deepened over many years.
In general, students must be brought to understand that AI is a powerful tool, but only a tool. In many ways, reading and writing can serve as a parallel. Just because a book or newspaper says something doesn’t mean that it’s true. And it certainly doesn’t mean that you can directly apply advice exactly as written. But knowing how to interpret and judge written information is an essential skill, as is the ability to express yourself in writing. It’s exactly the same with using AI and interpreting its output.
Finally, students should use AI in all aspects of their schoolwork. It’s not cheating to use AI to help write a paper; no more than it’s cheating to use a word processor with built-in spell-check. Using AI as an accelerant for both learning and task performance is the expectation for all high-productivity workers. In fact, staff who don’t use AI should be fired because of low productivity. Since we know that AI accelerates learning, students who don’t use AI to learn will progress too slowly through the curriculum. Worse, they’ll be acquiring and training useless skills. Requiring students to refrain from using AI for assignments is no different than requiring them to abstain from calculators and do all math in their heads. (It’s likely useful to perform a few exercises without modern tools, but the norm should be to do schoolwork with the same tools as adults use for all their work.)
Not only will students learn any topic better and faster if they are allowed to use AI for the lower-level tasks, but using AI for actual assignments is one of the best ways of gaining AI proficiency.
AI may soon become the best teacher of specific knowledge. In any case, one of the most important topics to learn (at any age) is how to best use AI. (Midjourney)
2: Influence Other Humans
No matter how AI evolves and how many tasks become easier to do with AI, the most fundamental essence of our existence is that we live in a world of other humans. Jean-Paul Sartre famously said that “Hell is other people,” but it would be more accurate to say that “the world is other people.”
Any success or achievement you may have in your professional career will come from dealing with other people. Write a report? Well, writing those words may seem to be your job, but the actual job is to impress the readers. Who are humans, even if an AI might be summarizing your report for the higher-ups.
To thrive in a world of other people, you must be good at influencing humans, so knowing what makes people tick is a core skill to acquire in an education. Since it’s guaranteed that many other people will be master manipulators, it’s also a core skill to learn how to identify and resist manipulation.
By “influencing,” I don’t mean lying, cheating, stealing, or being a con artist. Such behaviors will always be unethical, sometimes illegal, and most likely spotted by AI in the future.
Rather, I am referring to the use of social psychology findings, persuasive design guidelines, and behavioral economics to further your cause better than what will come from sticking to pure logic. Mr. Spock is a television character, not somebody who would be promoted to two stripes in any navy or space force.
Let’s take a simple example: the “liking principle” from social psychology says that people pay more attention to somebody whom they like and have a tendency to believe what this better-liked person says. Furthermore, the “halo principle” says that a positive impression in one domain transfers to other, unrelated domains. For example, being tall and good-looking means that people like you better (which may make a little sense) and, therefore, that they pay more attention to you (understandable) and tend to believe what you say (which makes absolutely no sense).
This is why people wear makeup. Thousands of years of history have shown that it works. And yes, men can wear makeup, too, and absolutely should do so on television, where they look horrible without it. This will be even more important as Zoom calls and other remote meetings move to 4K and higher resolution. Richard Nixon lost the U.S. Presidential election in 1960 because he thought it was unmanly to wear makeup for the televised debate. No candidate has ever made that mistake again.
Any time I have been on a professional television show, starting at a Danish equivalent of the Mickey Mouse Club at age 8, I have spent half an hour in the makeup room before going into the studio. (Actually, for the kids’ show, I think they only spent 10 minutes on makeup per kid because they needed to prep about 10 children for the show, and we didn’t have much patience for sitting around.)
Apply makeup before going on TV (or Zoom) to tame that shiny nose. (Ideogram)
Behavioral economics principles also drive many illogical behaviors. Saying that “Normally this product costs $10, but I can give you a discounted price of $8 that is only available today” applies two behavioral economics principles: anchoring (the normal price is $10, so that’s how you will value the product) and scarcity (the better price is only available today).
Some may think that persuasion principles, social psychology findings, and behavioral economics are unethical. But hundreds of thousands of years of evolution have made us this way: making fast decisions that obey these principles. They can certainly be abused. For example, false urgency is a common dark design pattern on websites that exploit the scarcity principle, maybe by falsely ticking down the number of remaining products available for purchase.
It’s a delicate balance to determine the ethics of influence techniques. If you falsely claim that something is about to sell out when it’s not, you’re unethical. But if you truthfully state that there’s only one copy left of something, you may be manipulating that customer, but in a way that doesn’t seem evil to me.
Any teaching of influence techniques should have 3 components:
influencing other people,
recognizing influencing techniques and resist them, and
staying on good ethical grounds and not abuse the principles.
It’s also valuable to teach the evolutionary psychology that brought the principles into this world: even though they lead to illogical behavior in the modern world, they were survival tools in the ancestral environment, which is why they got embedded in our genes. For example, in a starvation-level society depending on hunting, it would have been good for survival to believe tall and good-looking people because they had been eating more meat in the past and thus would be better hunters, making their hunting advice more likely to help you catch your own dinner. (Sadly, “tall and good-looking” doesn’t describe me. But I must admit that you would starve if you relied on hunting advice from me.)
Bottom line, the world is other people. This will remain so forever since AI is only a tool, and the other people are the system you must know how to work.
3: Build Projects
The final principle I expect to remain true no matter how good AI becomes is the ability to get things done. It’s amazing how many people are bad at executing. A future AI tool might be able to double or triple your productivity, but it’ll be a multiplier of whatever your base level of performance is.
So build. Get things done and out the door. Initially as hobby projects, and later as work products.
The habit of setting goals for yourself for accomplishing a project will almost certainly lead to success later in life. Also, you learn topics much better by using them for a real project instead of leaving them as dry textbook theory.
Build: do projects, initially for fun and later for learning. Building something that works beats any book learning or theory, and a can-do attitude will take anybody far. (Midjourney)
Other Useful Learning
The above three points are my main advice for future-proofing education: learn AI, learn to influence people, and keep building. Textbook learning has always been of questionable value, but with AI becoming a source of all knowledge, the only learning that will matter is the skills you can apply in practice to make things happen.
A few topics are sufficiently useful to be recommended, even if they fall into the textbook learning category: economics, business, and statistics. Understanding this trinity is the foundation for critical thinking and resilience. It’s less important to know exact formulas or even deep concepts like the central limit theorem.
In case your original statistics course didn’t emphasize useful knowledge, let me remind you that the central limit theorem states that the distribution of a mean approaches a normal distribution as the sample size increases, regardless of the shape of the population distribution from which the samples are drawn. This means that we can use the normal distribution as an approximation for the distribution of sample means, enabling straightforward statistical analyses and inference, even when the original data are not normally distributed.
The above one-paragraph definition of the theorem and its main practical application came straight from ChatGPT. (Of course, I checked before pasting, and edited a few words.) AI can give you all this info when or if you need it. But the ability to know in your guts how to relate to data is something you only pick up from experience applying statistics, economics, and business concepts to practical problems. Doing so is a twist on building.
Statistics, in particular, is so difficult and unintuitive that a course is the only way to learn. Unfortunately, most undergraduate stats classes are close to useless in teaching you how to apply statistical understanding to business and other practical problems. But you still need that course.
(See also my music video about the future of skills: YouTube, 3 min. song).
About the Author
Jakob Nielsen, Ph.D., is a usability pioneer with 41 years experience in UX and the Founder of UX Tigers. He founded the discount usability movement for fast and cheap iterative design, including heuristic evaluation and the 10 usability heuristics. He formulated the eponymous Jakob’s Law of the Internet User Experience. Named “the king of usability” by Internet Magazine, “the guru of Web page usability” by The New York Times, and “the next best thing to a true time machine” by USA Today.
Previously, Dr. Nielsen was a Sun Microsystems Distinguished Engineer and a Member of Research Staff at Bell Communications Research, the branch of Bell Labs owned by the Regional Bell Operating Companies. He is the author of 8 books, including the best-selling Designing Web Usability: The Practice of Simplicity (published in 22 languages), the foundational Usability Engineering (27,046 citations in Google Scholar), and the pioneering Hypertext and Hypermedia (published two years before the Web launched).
Dr. Nielsen holds 79 United States patents, mainly on making the Internet easier to use. He received the Lifetime Achievement Award for Human–Computer Interaction Practice from ACM SIGCHI and was named a “Titan of Human Factors” by the Human Factors and Ergonomics Society.
· Subscribe to Jakob’s newsletter to get the full text of new articles emailed to you as soon as they are published.
· Read: article about Jakob Nielsen’s career in UX
· Watch: Jakob Nielsen’s 41 years in UX (8 min. video)
What an insightful read!