Trust in the Age of AI

Good day, CharterFolk.

A fascinating week.

Atlantic writer Jonathan Chait, posting on X, cited the empathetic take I offered on Nikole Hannah-Jones’s piece in the New York Times, which sent thousands of new readers to the CharterFolk website.

My sense is that the NHJ piece might serve as something of a watershed.

This week we see another writer at The Atlantic make arguments that have been too long missing from the public discourse about education.

I had to write quickly on Sunday. I think I got most of it right. But my description of the two-pronged cognitive dissonance we feel parenting our own kids while trying to watch out for all kids wasn’t as precise as it could have been. I will return to that subject this weekend in a post I’ve been wanting to write for months.

For today, I want to expand the aperture to include a look at global forces at work in education.

Xi Jinping’s visit to the United States reminds me how illuminating broad international comparisons can be.

The trick is keeping the brush broad.

Go too specific and the individual bristles differ so much from one country to another that you miss what the paint is showing you.

I’ve written before about Spain, where a fully public system serving disproportionately lower-income students operates alongside a publicly subsidized concertada sector serving more affluent families, and a completely private sector serving families with even greater means.

And I’ve written about England, where over two decades policymakers have combined elements of strategies we associate with places as different as Mississippi and Indianapolis – stronger accountability alongside much greater school-level autonomy.

Whatever you think about these arrangements, things we discuss in the hypothetical in the United States are often made manifest in the literal elsewhere.

Helping us see more clearly what is happening here.

Which brings me to China.

For years, extraordinary PISA results from a limited number of Chinese provinces and municipalities helped create an impression of a system racing ahead educationally.

The broader national picture has always been more complicated.

And today it unfolds against a backdrop of startlingly high youth unemployment.

In August, China’s official unemployment rate for urban 16- to 24-year-olds, excluding students, reached 18.9%. That is approaching the extraordinary highs reported under the old series before the government stopped publishing it in 2023 and subsequently changed its methodology, resulting in a considerably lower reported rate when publication resumed.

Now comes The Economist with a story about another extraordinary development.

China’s crackdown on private tutoring.

Beginning in 2021, the Chinese government imposed sweeping restrictions on for-profit tutoring in core academic subjects. An industry once estimated at roughly $100 billion was devastated. Thousands of jobs disappeared. Major tutoring companies were decimated.

But something important didn’t disappear.

Parents’ desire to get their kids more instruction.

Human tutoring has consequently found ways to re-emerge – through one-on-one arrangements, relabeled offerings and other workarounds.

And now another alternative is arriving.

AI.

Which makes me wonder whether China’s crackdown on human tutoring might ultimately prove one of the greatest accelerants of AI tutoring happening anywhere in the world.

What do parents do when government makes it extraordinarily difficult for them to access the human tutoring they previously relied upon?

Increasingly, they can turn to AI tutoring in the privacy of their own homes.

If, in fact, anything remains private in our own homes anymore – a concern people in China might reasonably feel particularly acutely.

It takes me back to something I wrote here more than three years ago.

The biggest AI at play in education across the world, I suggested, might not be best understood to mean Artificial Intelligence.

It might mean Alternative Instruction.

That piece was about parents’ improving BATNA – their Best Alternative to a Negotiated Agreement.

What do parents do when the educational arrangement being offered to them simply isn’t acceptable?

They find another one.

In that 2023 piece, I focused on India, sometimes called a “tuition nation,” where enormous numbers of families pay for private instruction supplementing – and sometimes functionally supplanting – what their children receive in school.

I suggested that India might become a place where new models of alternative instruction would emerge and ultimately make their way around the world.

Some of that piece I got wrong.

BYJU, the gigantic tutoring company I featured prominently, subsequently collapsed.

But the phenomenon I was writing about didn’t.

If anything, this week’s Economist provides an even starker picture of why it hasn’t …

… describing a system that has expanded at breakneck speed without solving its fundamental quality problems.

At the top sits a relatively small number of excellent schools and universities, disproportionately accessed by the privileged. Below them stretches what The Economist describes as a “long tail” of poor institutions. At the school level, years of inadequate instruction leave enormous numbers of children far behind.

And families respond exactly as you would expect them to.

They look for alternatives.

Private schools in India now take in something on the order of $30 billion annually.

Coaching – tutoring – takes in more than $8 billion.

This isn’t some marginal educational supplement.

It is an enormous parallel system families have built because what government provides so often isn’t good enough.

And now AI is being layered on top of it.

What makes India particularly interesting today is how different its response appears from China’s.

Rather than principally trying to suppress alternative instruction, parts of India’s public sector are beginning to incorporate AI-enabled personalized support themselves in hopes of better competing with private providers.

In China, government responds to enormous parental demand for alternative instruction by restricting a massive private tutoring sector in pursuit of broader political and social objectives – and parents find ways around the restrictions.

In India, government responds to enormous parental demand for alternative instruction by reaching for the same new technologies families are discovering themselves, hoping to provide capabilities its existing public systems have never been able to deliver reliably at scale.

And then there is Kenya.

It is one thing to agonize over what AI will do in a landscape where families already have many alternatives.

It is another thing entirely where alternatives are scarce or clearly inadequate.

There, AI can offer something that may simply never have been available before:

Individualized instruction at enormous scale and extraordinarily low cost.

Arriving at a moment when Kenya itself is contemplating monumental changes to how education is organized and delivered.

And that brings me home.

For generations in the United States, affluent families have had a better educational BATNA than everyone else.

When school didn’t provide enough math, they bought more math.

When their child needed additional help reading, they hired a reading tutor.

When they wanted more test preparation, enrichment or individualized attention, they paid for it.

Recent research makes just how unequal this alternative-instruction market has been remarkably plain: nearly half of private-tutoring users in the United States come from families in the top income decile.

In that sense, the United States reflects a dynamic we see across the world – one level of educational opportunity available broadly, and another layer disproportionately accessed by families with the means to purchase more.

AI has the potential to scramble that arrangement.

It can make forms of individualized instruction that affluent families have purchased for generations available to vastly more families at dramatically lower cost.

Which raises an interesting question when school systems decide to restrict AI.

What happens when parents believe the decisions schools are making do not genuinely take into account the needs of their children?

Especially when those decisions seem to be coming from far away – from policymakers responding to broader political pressures rather than adults parents know and trust responding to the actual kids in front of them?

Increasingly, parents don’t have to accept those decisions.

They can route around them.

They can find Alternative Instruction.

Which brings me to charter schools.

The great advantage charter schools may have in this next era isn’t that we will necessarily be more pro-AI or anti-AI than anyone else.

Last week I visited one extraordinary charter school where students were learning to program an autonomous robot – sophisticated technology woven directly into the instructional experience.

Two days later, I visited another where an educator and her students stood proudly before a blackboard covered with a beautiful drawing done in colored chalk – part of a school culture that has made a deliberate choice to keep information technology out of the classroom based on its beliefs about how young children learn best and how human connection is fostered throughout a school community.

The contrast points to a more important truth.

The point isn’t that one of these schools has it right and the other has it wrong.

The point is that both are making thoughtful decisions about what is right for the kids and families they actually serve – and they have the agency they need to bring those decisions to life.

The world of better parent BATNA is upon us.

What matters most in that world is being understood to be making and enacting decisions close to kids.

Close to parents.

Close to communities.

Within healthy organizational cultures where unified and respectful educators – teachers and school principals – help parents grapple with incredibly vexing decisions in a world changing at almost unbelievable speed.

That, more than any particular position on AI, may be the great CharterFolk advantage.

Because the great educational question of the AI era may not be:

What policy regarding Artificial Intelligence should our system adopt?

But:

Who do parents trust to make decisions about the education of their children when many forms of Alternative Instruction are available to all?

And faster than many of us recognize, the arrival of better parent BATNA is making clear that our public positions regarding AI and education may have less to do with our visions for the future than with something much more immediate:

How much parents trust the schooling options they have for their kids right now.