What Algorithms
Can't Replace

I've been turning this essay over for a long time. Over the past two years I've spent hours with executives in my classrooms at Cheung Kong, talking about how AI is reshaping business. They speak with confidence about strategy, about pivots, about technology roadmaps. But every time the conversation turns to their own children, the tone changes. The certainty drops out. The analysis turns into a quiet sigh.

One of my students put it this way: "I can read where an industry is going over the next three years. But I have no idea how my child should navigate the next thirty."

That line stayed with me. In our professional lives we've trained ourselves to make sharp judgments. Yet with our own kids, those instincts fail us. Because we know something uncomfortable: the world is changing faster than anyone can plan for.

The World Economic Forum's Future of Jobs Report 2025 projects that 92 million jobs will be displaced globally by 2030.[1] MIT researchers have found that AI is already capable of replacing nearly 12% of the U.S. workforce.[2] Anthropic's CEO Dario Amodei has said AI will be writing 90% of code within three to six months.[3] Data entry, translation, junior coding, financial analysis — all the things we pushed our kids to master — are being quietly taken over.

So what can't be replaced?

After months of reading and arguing and changing my mind, the answer has grown clear. It isn't a particular skill. It's three deeper qualities — passion, empathy, and taste — which lead, in turn, to creativity, connection, and judgment. These are the three areas where AI is weakest. And they are the three things the standardized education system is least equipped to grow.

IPassion → Creativity

Let me start with a story you may know.

At age five, Einstein was made to study violin. He found it unbearable — once, he threw a chair at his teacher. Most parents today would have given up, or anxiously swapped instruments. But his mother Pauline did neither. She was a pianist; music was always in the air at home. She didn't push, and she didn't quit. She just let the music stay.

Eight years later, when he was thirteen, Einstein happened to hear a Mozart violin sonata. Something caught fire. From that moment on, he never put the violin down. He said later: "If I were not a physicist, I would probably be a musician. I often think in music." About relativity, he was more direct: "It occurred to me by intuition, and music was the driving force behind that intuition."[4]

A boy his teachers once called slow was given eight patient years by his mother — long enough for his own spark to find him.

This isn't a feel-good story; there's research behind it. Teresa Amabile at Harvard Business School has spent thirty years on creativity, and her findings keep pointing to the same thing: people are most creative when motivation comes from genuine love of the work itself, not from rewards or pressure.[5] Analyzing 12,000 employee diary entries, she found that a person's mood on a given day could even predict their creative output that day and the next.

Mihaly Csikszentmihalyi — the psychologist who coined flow — interviewed 91 exceptional creators and found something surprising: most of them were not child prodigies. What set them apart wasn't raw talent. It was intense curiosity.[6] More striking still: the single most important value these creators said their parents had passed on wasn't discipline or hard work. It was honesty.

Why honesty? Because in an honest household, a child doesn't have to pretend. She can say "I'm not interested in this" — or, just as freely, "this is amazing" — without calibration. Honesty is the soil passion grows in. A child who is always reading her parents' expectations, always performing the right interests, rarely finds her own spark.

Now back to AI. It can generate a polished essay, a perfect image, a clean block of code. But it has no passion, no "I have to do this" pulling at it. The Oxford Institute for Ethics in AI points out that generative AI lacks two essentials of creativity: intentionality and critical self-evaluation.[7] AI doesn't wake up at two in the morning because of an idea.

So the first thing I want to say to parents is this: don't rush to correct your child's "useless" hobbies. If your kid is obsessed with building things, raising insects, making up stories, or taking toys apart — that obsession may be the strongest armor they will ever have. Our job isn't to pick the right path for them. It's to do what Einstein's mother did: keep the space for that fire open, and wait.

Ken Robinson, whose TED talk has been seen over a hundred million times, put it sharply: we don't grow into creativity; we're educated out of it.[8] Standardized schooling was built for the industrial era, and it trains exactly the skills AI does best: repetition, compliance, measurable output. In the age of AI, that observation lands with a sting.

IIEmpathy → Connection

The second story is one I've never been able to shake.

In Canada, there's a program called Roots of Empathy, founded by Mary Gordon. The idea is simple: once a month, a mother brings her baby into a classroom, and the children sit in a circle and watch. They notice when the baby smiles, when she cries, when she looks frightened.[9]

One of the kids in the program was a boy named Darren. He had lost his mother at four and had been moved between foster homes ever since. On this particular day, he walked up to the front and asked to hold the baby. He cradled her, rocked her for a moment, and then asked, almost in a whisper:

The first time I read that, I sat at my desk for a long time. A boy who had never been properly loved still longed to love someone else. Empathy isn't something you teach. It's something that wakes up inside a child, slowly, as he is seen, heard, and accepted.

The program has reached nearly a million children in 14 countries. Multiple randomized controlled trials have shown that it significantly increases empathy and prosocial behavior, reduces aggression and bullying, and the effects hold up one to three years after the program ends.

Why does empathy matter more in the age of AI?

Because of a counterintuitive reality: the more advanced our technology, the lonelier we get. In 2023, U.S. Surgeon General Vivek Murthy declared loneliness a public health epidemic — about half of American adults report feeling lonely, and chronic loneliness carries a health risk comparable to smoking 15 cigarettes a day.[10] Globally, more than a billion people say they feel lonely.

Nature published a striking study this year: people rate AI-generated empathetic responses as "higher quality and more effective" than human ones — and yet they still overwhelmingly prefer empathy from another human being.[11] We don't want the right words. We want someone, somewhere, actually cares about me. An algorithm cannot give that.

Daniel Goleman's research on emotional intelligence backs this up with hard numbers. TalentSmart has found that 90% of top performers have high EQ, and EQ accounts for 58% of job performance across occupations.[12] Any experienced leader knows this in her bones — how many of your biggest breakthroughs came down to a single honest conversation, a single moment of trust?

How do you raise empathetic kids? Research points to several paths. A study in Science showed that reading literary fiction measurably increases empathy.[13] UCLA researchers found that just five days of high-quality, face-to-face interaction — no screens — was enough to significantly improve children's ability to read others' emotions.[14]

But more important than any technique is a reframe: screens aren't the enemy. The absence of real human presence is. Screens are now basic tools for learning and living; there's no point fighting that. What we can do is protect the irreplaceable moments. Put the phone down at dinner and actually talk. Curl up together for a bedtime story. Head outside on a Saturday with no plan at all. These "wasted" hours are where children learn the one thing that matters most — how to actually be present with another person.

IIITaste → Judgment

The third is the one I feel most personally.

In 1972, Steve Jobs sat in on a calligraphy class at Reed College — serifs and sans-serifs, the spacing between letters. At the time, he had no idea what any of it would be good for. Ten years later, when he was designing the Macintosh, all of that feeling for beauty came back. In his 2005 Stanford commencement address, he called calligraphy "beautiful, historical, artistically subtle in a way that science can't capture."[15]

And he said something I have come back to many times:

"Ultimately it comes down to taste. It comes down to trying to expose yourself to the best things that humans have done, and then try to bring those things into what you're doing."

In 2026, that sentence carries more weight than when it was first said. AI can generate an essay, a poster, or a song in a second. The cost of creation has dropped to zero. Which is precisely why Eric De Castro, in a widely-shared essay this year, argues that the internet is being drowned in gray slop — AI-generated content that's technically fine but spiritually empty. His conclusion is sharp:

Any entrepreneur knows the word moat. We talk about brand moats, tech moats, network-effect moats. But as AI flattens one technical advantage after another, the last real moat left may be taste itself — knowing what's good, what's worth doing, and what to refuse.

Taste sounds soft, but at its core it's judgment. Over a century ago, the Chinese educator Cai Yuanpei saw this clearly. Quoting Schiller, he wrote: "To make the sensuous person rational, there is no other path than first making him aesthetic."[17] In an age flooded with AI-generated content, that observation is almost prophetic. To keep your bearings in the information torrent, you first need a feel for what's beautiful and true.

Research in Frontiers in Education this year confirms the link: there's a strong positive correlation between critical thinking and the ability to identify misinformation.[18] The capacity to tell real from fake, good from bad, rests on the same foundation as taste.

So how do you cultivate it? Many parents I talk to immediately think: sign them up for an art class. But taste isn't technique. It's the slow accumulation of exposure to good things, until it becomes an instinct. Take your child to museums, to concerts, to good books, let them stare at clouds in the woods. These are all paths. But in a world saturated with screens, there's one thing that matters even more: help your child become a curator of content, not a consumer of algorithms.

Recommendation algorithms are optimized for time-on-screen, not for the growth of taste. They reward what's popular and provocative, not what's deep or meaningful. As De Castro puts it bluntly: you can't develop taste by eating junk food. The answer isn't to ban screens — that's neither realistic nor necessary. The answer is to teach kids to search actively instead of scrolling passively, to save things instead of swiping past, to ask "why do I like this?" instead of just reaching for the next clip.

Jobs said something else worth keeping in front of any grown-up who makes things: "Technology alone is not enough — it's technology married with the humanities that yields the results that make our hearts sing." The people he hired, he said, were "musicians, poets, and artists who also happened to be the best computer scientists in the world." Most people don't have enough varied experiences to connect dots across, so the solutions they generate end up being very linear.[15]

That's true for raising kids as much as for running a company.

IVThe Hardest Homework is Our Own

I have to be honest about something. All three of these qualities — passion, empathy, taste — share one feature. None of them can be drilled into a child.

Passion needs room. Empathy needs to be witnessed. Taste needs unhurried time. And all three run directly against what our generation is best at: planning, optimizing, squeezing out inefficiency.

Thirty years of Amabile's research keeps landing on the same point: external pressure and over-control are the greatest enemies of creativity.[5] If a child already has intrinsic motivation for something, adding material rewards can actually weaken her love for it — what psychologists call the overjustification effect, and it's been confirmed in experiment after experiment.

A paper from the IZA Institute of Labor Economics puts it more starkly:

That's an uncomfortable sentence for most driven parents. We're good at doing something about it — see a problem, solve it; see a gap, close it; see a blank, fill it. But a Chinese educator named He Lingfeng once said something that stopped me: "Good parenting means learning to leave blank space in a life. How well our kids turn out isn't always because of what we did — it's sometimes precisely because of what we didn't do."

The great translator Fu Lei raised his son Fu Cong by this philosophy: first a person, then an artist, then a musician, and finally a pianist. He put character above talent, being a human being above doing great work. And yet, late in life, he confessed his earlier harshness: "My child, I mistreated you. I will never be able to make it up to you… It's pathetic that my fatherhood didn't truly wake up until I was past forty-five." On Fu Cong's gravestone is a single line from his father: "A pure heart, when it is alone, creates a world."

I think often about how much complicated love is compressed into that sentence.

A longitudinal study of 148 exceptional young adults found that six childhood traits predicted success: independence, strong will, cooperativeness, moral discernment, choosing good friends, and integrity first. Not one of them is something a standardized test measures.

I know the anxiety is real. We're living through a transformation none of us has seen before — AI is advancing faster than anyone predicted, our own industries are being upended, and that's before we think about the world our kids will face in twenty years.

And here is the line I wish I had opened the essay with, instead of closing with it: AI does not make people obsolete. AI makes more of whoever a person already was. If our kid was building, AI helps them build at ten times the scale. If our kid was scrolling, AI helps them scroll at ten times the speed. The machine has no preference between those two outcomes. Only the child does — and only if we have helped them grow the inside that decides.

But maybe because the future is unpredictable, we should return to what lies underneath it. Passion lets a child, in an age when AI can do almost anything, still do the things AI cannot — because real creation is never just craft. It comes from lived experience, curiosity, and love. Empathy lets a child, in an age when algorithms connect everything but leave us lonelier than ever, truly reach another human being. The rarer that ability becomes, the more valuable it is. Taste lets a child, in an age when gray slop is drowning the world, still tell what is good, what is real, and what is worth doing.

The path to cultivating all three is strikingly similar. None of them can be taught directly. They grow, naturally, in an environment of love, freedom, and rich experience.

In a world where AI can write most of the code and replace half of white-collar jobs, the ultimate goal of education is no longer to raise a useful child. The useful skills are being taken over, one by one. The goal is to raise a whole child: someone with passion in her heart, other people in her sight, and enough taste to tell what is true and beautiful in this world.

A child like that will not be obsolete in any era.