I was asked to be on a panel this week about innovation in education. Here’s some thoughts:

public education for an educated public
A great sticker I saw somewhere.

Hi everyone – glad to be here.

Our education system, which maybe made sense in the ’90s… no longer does today.

First, most students can’t access a high quality education. A billion students globally – including many in the US – are hungry to learn but don’t have access to a good in-classroom education. That’s both college and K-12.

Second – the cost is way too high. Not just tuition and student debt, but also the carbon cost – all those millions of college students flying around to small handcrafted schools where they may not be learning that much anyway. One cross-country flight is (literally) a ton of carbon emissions. International is more.

The good news is we have a lot of technology that can help. But we need the willpower to change the system. And we need to believe in the upside of doing it, and really, actually leaving no child behind.

There’s also enough money to do it, but it needs to be reallocated.

Education globally is $5 Trillion a year. Almost all that goes to buildings and people… very little to educational content, which is basically a commodity. Textbook companies were never big businesses.

So entrepreneurs haven’t really ever been drawn to education, except for those working on starting new types of schools.

That changed twenty years ago, with bootstrap Internet entrepreneurs like Sal Khan creating learning content on YouTube, or study apps, or online courses targeted at professional skills and certifications.

Now that hundreds of millions of users are using those resources, big money and talent is piling into consumer online learning.

And we have an innovation pipeline, which I’ll describe in two categories:

One – and the one I’m most excited about – is innovation supporting online learning with human teachers. There’s the small-class model with a teacher on Zoom. There’s the Peloton model, with a teacher and 20,000 students. There’s the tutoring-on-demand marketplace model. There’s lots of great human-centric online learning models.

The other which I have mixed feelings about, is innovation to get computers to do some or all of what human teachers currently do. Like self-driving cars, nobody asked my opinion on this and its happening anyway. 

We already had apps like Duolingo which can teach an intro language course. Now we’re seeing evidence that AI, coupled with human curation and a massive amount of data, can do even more…  potentially guiding students through a subject like biology from practice to mastery using human-created content, without humans being present.

So which of these models will win the online learning race?

I don’t know, but I think COVID-19 has fired the starting gun.

I’ll leave you with two thoughts on access and costs.

On access, technology can provide some really great learning experiences for extremely low or no cost (the no-humans model has zero variable cost). This could dramatically democratize access and improve the educations of hundreds of millions of students.

But even online with humans teachers involved, how do we replace the face-to-face value prop? The inspiration, the socialization, the babysitting?

This can be figured out – we’re already solving lots of other hard online problems like privacy, bullying and cheating – but we cant just leave this to the market so only the rich kids get inspiration and socialization, and everyone else just gets babysitting.

On cost – who’s going to challenge the status quo first? I’m secretly hoping a food fight breaks out among major universities now that COVID-19 has normalized online courses.

Cal State recently told its 400,000 students they’ll be online-only this fall. Couldn’t Stanford or MIT or UCLA say to those students, “Hey, pay us the same $10k and get a better online degree from us instead?” That four billion a year in tuition by the way… just the Cal State students.

Is there any reason Stanford couldn’t have two million graduates a year, rather than two thousand?

I’ll leave it there… look forward to hearing your thoughts on all this.