Why the ‘education crisis’ is a global malady (and how AI can fix it)
Let’s climb onto the shoulders of the AI-in-education pioneers Ethan Mollick and Sal Khan to radically reimagine a new, “post-linearalist” ecology of learning
Part One
As the purpose of this series is to share a few thoughts on learning and artificial intelligence, let’s start with an educational exercise:
First, pick any country in the world. Recall the dreamy destination of your honeymoon. Pick a spot off your bucket list. Throw a dart at a map. Then, Google the name of that country plus the search term “education crisis”.
Toying with this idea myself, I picked Bhutan off the top of my head as my wife, Debra, really loved her visit there along with her sister, Denise. Sure enough, teachers in Bhutan, which has been named the happiest place on Earth, are unfortunately fleeing the profession. Test scores are dropping, STEM instruction is mediocre at best, and children’s futures are at risk. Sound familiar? Being more intentional, I also searched Finland — often cited as the country with the world’s best schools. The math and reading skills of Finnish children are plummeting, according to a government report issued just a few months ago. You can get carried away: In Hong Kong, 6,500 teachers quit last year. In New Zealand, there is a student brain drain. In Denmark, another place where STEM is stuck. In China, a ban on tutoring has millions of parents panicking. In Egypt, parents are sacrificing food to pay for tutors.
To be sure, the picture is not universally grim. As just one example from a favorite weekly newsletter on good news, Future Crunch, in Nepal the net enrollment rate for high schoolers in poor districts has doubled in the last five years to nearly 50%, and overall the graduation rate for girls now exceeds boys for the first time. Yet my argument remains that the challenge is global and we need urgent, robust, and comprehensive embrace of AI in education. I even sought a bit of help on the case from Google’s Bard as well since it is trained on the latest data available on the Web: “According to my knowledge,” came the anthropomorphic response, “there is no country in the world that is not facing some kind of education crisis.”
My point here is not to suggest that we walk away from the hotly debated local issues of school choice, teacher salaries, crowded classrooms, stifling bureaucracies, pandemic recovery, testing methodologies, curriculum bans, cancel culture, and so much more. With our son just having started 8th grade and our daughter starting her freshman year of college, trust me… I share the concerns of all parents. Deeply.
But I do suggest that we should also consider these problems — as close and immediate as they are — as smoke alarms alerting us to an even larger and global fire that we need to worry about. That fire is in fact a planetary learning disease. We need to examine this burning forest if we’re really going to do something about our own smoldering trees.
Linearalism in our nonlinear reality
Now before I get too far along, I want to introduce the educational AI pioneers I alluded to above. Ethan Mollick is a professor at my MBA alma mater, The Wharton School at the University of Pennsylvania. At the joint Arizona State University — Global Silicon Valley Summit, he described how learning outcomes in his classes have skyrocketed while tedium has diminished as he’s required students to use and master AI tools. Mollick’s post on the coming “Homework Apocalypse” is a good place to begin acquaintance with his marvelous work. Sal Khan, founder of the non-profit Khan Academy, recently shared at the annual TED conference his vision to provide every student in the world a personal tutor and every teacher an assistant — for free. Sal’s talk was my favorite at this year’s TED because of its incredible potential for both students and teachers globally. Pano Kanelos, founding president of the new University of Austin, is another name to add to the list of learning disease fighters. His reinvention of higher education is something I wrote about in April. He recently asked me to join his AI Advisory Board, and I gladly accepted.
I’ll return to these pioneers and a few others in my next essay that will follow. My destination in this series is the new “learning ecology” they are creating.
But before that, let’s establish the outlines of this “disease”. Why is it flaring up everywhere from Austin to Auckland to Ankara? In our analytical discourse, it’s a concept that lacks a name. So I’ll give it one, not yet in the Oxford English Dictionary: “linearalism.” Of course the deepening crises from Uruguay to Uzbekistan all have unique dimensions, but the common thread is linearalism — the one-step-at-a-time process that defines the way most of our modern institutions work, and for that matter the way we tend to think.
Linearalism is not confined to education, of course. Since the 16th Century, our societal operating system — our civilizational “OS” as it were — has really been defined by the profound (but profoundly outdated) thought paradigms of Newton and Descartes. These thinkers focused us on the mechanistic — that which can be physically measured, publicly observed, and repeatedly tested. No complaints there. Linearalism has given us modernity, with tools and advances from double-entry bookkeeping to the Dewey Decimal System to the scientific method itself, and thus all the fruits of technology that we take for granted.
But this is the world we’re leaving in the digital dust with non-linear complexity surrounding us in the forms of climate change, an algorithmically-driven economy, emerging zoonotic disease, and 3D-printed homes, prosthetic limbs, and musical instruments. Add to that the atomization of all forms of media, and an explosion of heterogeneous data challenging, and even confounding, all of our institutions. We’ve all heard the famous estimate by former Google CEO Eric Schmidt that all the information humanity created from the dawn of our species until 2003 is equal to the volume of data we now create in two days. But AI is so much more than a lineralist stage in that process. Rather, AI is the convergence of the internet, data cloud, mobile connectivity, blockchain, mRNA vaccines, and it in fact represents what I’ve called history’s “Fourth Surge” of data that will shape the 21st century.
The transition to a networked, digital, hyperconnected, non-linear reality has been characterized by the technology sage Kevin Kelly as a planetary “phase transition”, a term borrowed from physics and about which I wrote in a series two years ago on the evolution of data. When ice melts from solid to liquid, the molecule H2O undergoes a phase transition. Then when heated to boiling, water becomes a vapor. By analogy, our very civilization is moving from this analog Newtonian reality, where outputs can be measured proportionally to inputs, to a quantum reality where all linearalist bets are off. In a linearalist world, much is complicated. But in the emerging non-linear, quantum reality that we are just beginning to understand, complexity rules.
“Man-made complex systems tend to develop cascades and runaway chains of reaction that decrease, even eliminate, predictability and cause outsize events,” wrote mathematician and thinker Nicholas Nassim Taleb in his seminal book Anti-Fragile — Things That Gain from Disorder. “So the modern world may be increasing in technological knowledge, but, paradoxically, it is making things a lot more unpredictable.”
No institution — from Lehman Brothers to the World Health Organization to your local city council — can escape Taleb’s insight. But education — or learning as I prefer to think about this — is, without a doubt, the keystone institution on which all others depend to cope with the coming change.
Our education/learning systems, whether in Bhutan or Finland, are products of the 16th century invention of sovereign nation states at the Treaty of Westphalia ending a long series of European wars. Among the results was the need for uniformly acculturated “citizens’’. This rough model was then refined by the demands of the 18th century industrial revolution that needed workforces with uniform and standardized skills. Fast forward and we have the familiar linearities of education, like A to F grading and K-12 classrooms, divided between primary and secondary schooling. There’s a tertiary stage of universities organized as undergraduate, graduate, and postgraduate stages, and all of this runs on a currency of accumulating “credits”. It’s all overseen by professors whose careers are defined by a three-step “track” that ends with tenured professorism.
Consider that this conveyor belt model of schooling is organized pretty much along the same lines as Henry Ford’s famous assembly line installed in 1913. So check out Ford’s use today of so-called “rapid manufacturing processes”, which include 3D printing and the cousin technologies of fused deposition modeling and laser sintering. Teslas are made with what are essentially distributed, miniaturized assembly lines that robotically complete so-called “modules” for the powertrain, interior, body, and other component combinations that are then snapped together like Lego bricks. I find inspiration in this for rethinking the way we enable learning.
To be sure, education’s AI reformers like Mollick and Khan stand on the shoulders of many giants who have sought to demassify the assembly line approach to schooling over the decades — even if they’ve not consciously framed their work in the terms I am using here. Reforms and innovations that have challenged linearialism in our schooling would include the heroic founding of the first university “extension” for lifelong learning in 1892 at the University of Wisconsin — Madison, or the so-called “Montessori movement” of child-centered education, which began in Italy in 1907. Just a few alumni of Montessori schools include late First Lady Jacqueline Kennedy Onassis, actor George Clooney, Jewish diarist Anne Frank, and the tech titans Mark Zuckerberg, Jeff Bezos, Larry Page, Sergey Brin, and Bill Gates.
There is the Head Start Program for early childhood education targeting low-income families begun in the 1960s, whose grads include comedian Chris Rock and Ford Foundation president Darren Walker. On the opposite side of the planet, the world’s first virtual university, created in India in 1985, Indira Gandhi Open University, now teaches more than seven million students in 34 languages. Or there’s the idea hatched 30 years ago in Minnesota, so-called “charter schools”, which have grown to number more than 7,000 across the country.
Countless improvements to our basic linearalist model to impart knowledge have improved the lives of millions. So let’s honor and turbocharge this century-long struggle by so many to move from one-dimensional, conveyor belt education to holistic and three-dimensional learning.
To further the goal of a true ecology of learning, we are at a juncture where access to learning has never been easier. In my book published last year, The Entrepreneur’s Essentials, I emphasized the importance of the means to avail ourselves of books, podcasts, online and offline seminars and conferences, and all the tools of virtual learning in Chapter 4 — The Importance of an Always Be Learning Life. I led by example here by giving my own book away for free online, in addition to it being available for purchase at Amazon.com (with all proceeds going to benefit the Kendra Scott Women’s Entrepreneurial Leadership Institute at the University of Texas at Austin.)
But that’s not enough. Pondering this idea of learning as a planetary problem in recent days, and discussing it broadly with an array of friends and colleagues, my mind wandered back to a book I read more than a quarter century ago as an undergrad at UT Austin. The book itself was written a quarter century before that: Future Shock by the late seer Alvin Toffler. His remains a powerful call to integrate education into the new “OS” of humanity.
“The illiterate of the 21st century will not be those who cannot read and write,” Toffler wrote, “but those who cannot learn, unlearn, and relearn.”
It’s an insight that has aged well. Allow me to build on this idea and to connect it to our fast evolving AI adaptation generally, and deployment of AI in schooling specifically. The case I want to make is that we err when we consider this radical technology outside of the history that began with the development of the transistor in 1947. The dawn of the era popularized by Toffler as the “information age”– our post-industrial society.
In that book and in other writing over the years, including the classic compendium of 19 authors which he edited, Learning for Tomorrow — the Role of the Future in Education, Toffler argued presciently that as we have moved to a society of information primacy, data-driven businesses, decentralization, and distributed production. Yet our education/learning system has remained a relic, mired in a model of mass uniformity and standardization.
“Like the elders of the tribe living on the riverbank, they (our political leaders) blindly assume that the main features of the present social system will extend indefinitely into the future,” Toffler wrote. “And most educators, including most of those who regard themselves as agents of change, unthinkingly accept this myth.” Nailed it.
OK, Toffler’s line is unfair to the tribal elders. We now know they were wise in ways we should emulate, even contributing to the deliberations of America’s founding fathers, with insights that helped shape the Declaration of Independence. This idea was brilliantly explored by the anthropologist/archeologist team of David Graeber and David Wengrow. But considering that Toffler wrote that a half century ago, let’s give him a pass on the elders and agree that his far-sighted warning is as timely as the latest version of ChatGPT.
No longer can we ignore the wisdom of Toffler, Kelly, Taleb, and so many others who have been warning us for decades now that our institutions — particularly our learning institutions — have fallen far behind our evolution as a species that is extending and expanding our consciousness with technology. As I suggested above, the seemingly sudden emergence of these new tools of AI, based on so-called Large Language Models or “LLMs”, have not so much ushered in a new era as they are consolidating it. It’s the warp speed acceleration of Moore’s Law, and the fusion of the output of all the information technology since that transistor was invented in 1947. Yes, our linearalist brains struggle to intuitively understand the exponential change, the new civilizational OS, that is upon us. But our primary hope — perhaps our only hope — is that a new OS will be nurtured through a learning system that is at once exponential, digitally-driven, and geographically unbounded.
We got an early glimpse of this in the pandemic, and ironically not just the example most often cited of Zoom-based classes. For sure, schools’ migration to virtual space, however imperfect, was somewhat adequate for those students with good internet access and supportive home lives. For those without, we failed them. But imagine what learning would have looked like had we endured a similar virus in 1990, the year I started my college journey at U.T. Austin. The current debate over the pandemic “hangover” of lost learning opportunity is a large topic for another day.
But less discussed when our attention turns to Zoom is the learning and skills that enabled a large swath of the workforce to go remote in the first place. Before the pandemic, just 17 percent of the U.S. workforce worked remotely. At the height of the pandemic, almost half the workforce was doing so remotely. Among those remote workers was the global team of science-in-the-fast-lane researchers who delivered us a vaccine in less than a year. We certainly transitioned successfully at my company data.world, which like many other now-hybrid tech companies is not returning to a full-time “in-office” operation.
Distributed decision-making saves the planet
But where did we get the skills, and not just the technical skills, to move almost overnight into a virtual workspace and to do so successfully? The largely unsung hero of the pandemic is the gaming industry that has taught at least two generations the skills of distributed decision-making. After all, what is StarCraft (1998), which teaches strategy, or Minecraft (2009) that teaches players how to build anything they can imagine?
Some 41% of the world’s population plays online games, some two-thirds of Americans do so, and more than 80% of the world’s population with access to the internet — a total of 3.2 billion people on networked keyboards, mice, and joysticks. For the most part, gaming is seen by educators as a distraction at best, incapacitating at worst. But as Jane McGonigal, author of Reality is Broken — Why Games Make Us Better, argued at Austin’s SXSW festival a few years ago, change is nigh. One 6–12 grade school in New York City, she notes, Quest to Learn, is the nation’s first public school built atop a game-based curriculum.
I would argue (and I’m sure to be challenged on this) that the most important educational advance of the late 20th and early 21st centuries has been scarcely considered by any college, department of education, or school board. Although I want to give a shout-out to Wharton for Markstrat, a game I really enjoyed playing as one of our multi-week MBA assignments in class back in 1998. For it was the rise of gaming through which we all learned from one another, taught one another, and communicated across time zones and cultural boundaries. This global experiment in immersive education gave us these skills of distributed decision-making that helped save the economy during the pandemic and are now at the heart of the remote work paradigm and a permanent fixture of the world’s workforce.
And now all of this — from that first university extension, to India’s seven million university students in a virtual school, to the Quest to Learn high school in Manhattan — is converging with the AI disruption of our centuries-old, linearalist system of schooling. This is the ultimate disruption of education’s planetary polycrisis.
In Part Two, I explore how this new learning ecology is evolving, from the immersiveness of distributed decision-making skills enabled by gaming to the revolutionary ways AI in the classroom is challenging all of our assumptions about the nature and mission of schooling. It’s a time to be bold and challenge ourselves to really solve our globally failing educational system.
