By Sal Khan, Founder and CEO of Khan Academy
We’re seeing more and more schools succeed by leveraging technology thoughtfully. But that success is never about the technology alone. The teachers and administrators deserve most of the credit. And often these technology-enabled schools have other advantages that are hard to replicate. That’s not just a question of giving credit where it’s due. It matters because our real goal is bringing these benefits to classrooms everywhere, and we can’t do that if we underappreciate all that has to happen to see notable success. Credit the technology too heavily and you set expectations that are far more likely to end in failure than in success.
Schools that succeed
I’ve spent much of the past 15 years championing a simple idea: students can learn more, faster, by leveraging technology to allow them to master concepts at their own pace. The intent has never been to replace teachers or teaching but, instead, to free up time in class for more active human interaction. With more time, teachers can lead deeper human-centric Socratic dialogues, group simulations, and peer-to-peer interactions.
The intent has never been to replace teachers or teaching but, instead, to free up time in class for more active human interaction.
– Sal Khan
I’ve believed this idea so strongly that 13 years ago I co-founded a nonprofit school built around it: Khan Lab School. Years later, I helped launch Khan World School, a free online charter school run in collaboration with Arizona State University. In both cases, the results have been extremely strong.
At Khan Lab School, we do not select students based on IQ or test scores. Instead, through Khan Academy and deep, hands-on experience, we leverage personalized, mastery learning and have high expectations of our students. High school math, science, and humanities are typically mastered by the end of 8th grade. All of our students take college courses in 9th grade. The median SAT score for our recent graduating class was 1520 (99th percentile). Our motto is “Everyone is a student, everyone is a teacher.” We routinely have upperclassmen teach courses like philosophy, multivariable calculus, or linear algebra—courses that aren’t usually offered at most high schools and are normally taken in students’ sophomore or junior year of college. In addition, our students also have time to pursue their own interests. One built a business for guinea pig enclosures. Within a year, it had hit half a million dollars in revenue and was profiled on CNBC. Another student built an advanced math course on Khan Academy that has been used by millions of learners worldwide. In 2024, one of our students won the International Linguistics Olympiad.
At Khan World School, which is free to any student in Arizona, the inaugural class of roughly 50 9th graders posted average scaled-score growth of 2.8 to 4.6 times the typical year’s progress in math, more than three times the typical growth in reading, and more than five times the typical growth in language arts, depending on the assessment window. Steve Levitt, the University of Chicago economist behind Freakonomics, has been candid about his own initial reaction to those numbers: When the school first sent them to him, he assumed there had been some kind of error and said he’d never seen results like that from a school in 25 years of education research. He insisted on getting the raw data himself before he’d believe it. He was ultimately convinced enough to co-found an in-person high school in Arizona directly based on the Khan World School model.
Acknowledging the caveats
I’ve helped start these schools because I want to first see what is possible when we build schools based on the principles of thoughtful use of technology, student agency, personalization, and active learning, and then share any insights we get broadly.
But we intentionally haven’t put a spotlight on these schools as prime examples of what can be done everywhere just with the technology alone. Yes, they are heavily leveraging Khan Academy, but this is still a small fraction of the school day. Most of the day is spent in discussions, reading, writing, and doing projects. Khan Lab School is based in Silicon Valley and draws students and families who are very invested. We’ve tried to keep the costs comparable to those of local public schools, but tuition is still in the mid-$30,000 range to cover six-figure teacher salaries and high real estate costs. Because of its association with Khan Academy, we’ve had everyone from Malala to Mark Rober to Francis Ford Coppola as guest speakers. Few schools get those kinds of speakers, and I don’t want to pretend that’s replicable.
We intentionally haven’t put a spotlight on these schools as prime examples of what can be done everywhere just with the technology alone.
– Sal Khan
Khan World School at Arizona State University is much more accessible; it is free to any resident of Arizona and costs $10,000 per year for anyone who lives outside the state. Families and students still self-select into the program, and we tend to attract students who are understimulated academically at traditional schools. As much as Khan Academy technology has played a role here, so has the teacher-led Socratic dialog, the high volume of reading the students do, and the Oxford-style advisor check-ins that the school provides.
I can’t fully separate out how much of our results come from our model versus the advantages of resources and self-selection. But I am not aware of any schools—even private schools charging more than $60,000 per year with strong resources, self-selection, and admission filters—that have these levels of academic gains. And this is before even considering other benefits, like students having the time and freedom to pursue their passions. Anecdotally, it seems that students in our schools feel more intrinsically motivated and less stressed than their peers in surrounding schools. This is why we are seeing an increase in demand for access to Khan Lab School and Khan World School, and we are carefully thinking through how to scale the models.
Results in a large public school district
With that said, I can appreciate why folks would still be skeptical of outsized results at these schools. That’s exactly why our results in Newark Public Schools matter. Newark Public Schools in New Jersey is a large, complex public school district in which we’ve seen gains on the state math test scores that are three times the state average. They’ve done this without any of the benefits of self-selection or unusual resources. There are edtech watchers who sometimes cite the “5% problem,” which refers to cases in which a small minority of students actually engage enough with the technology to get benefit. But in schools in the North Ward of Newark, we are seeing nearly 70% of students engage deeply, and some schools are seeing even a higher percentage of students than this. So the difference between 5% and 70% is the level of thoughtfulness and commitment from school leaders on this implementation, not an on-average difference in the underlying technology or students.
Perhaps the most compelling proof that technology can dramatically accelerate outcomes when well supported by humans is a large, independent, randomized trial involving government schools serving low-income students in Uttar Pradesh, India. In that trial, schools that only had access to Khan Academy logged average practice times of a handful of minutes a week. But schools that also included a staff member whose full-time job was to make sure the platform was actually used well saw students practicing more than six times as much. They also posted math gains equivalent to two to three years of typical schooling in that setting. Same software. Same students, on average. That’s the whole argument in miniature: the technology matters, but only when it’s strategically implemented with the added support of teachers and administrators.
The technology matters, but only when it’s strategically implemented with the added support of teachers and administrators.
– Sal Khan
The people who make it work
What is consistent across all four scenarios—India, Newark, Khan World School, and Khan Lab School—is that the results didn’t come from technology alone. They came from technology paired with exceptional leadership, sustained teacher training, and disciplined curricular choices carried out over years.
That’s the pattern I’d want anyone to internalize who is looking to accelerate learning by leveraging technology. It’s a real thing. The results can be dramatic. But it isn’t the technology doing that on its own. Some students pick up tools entirely on their own and go remarkably far. A young woman in Afghanistan used Khan Academy as her only school when the Taliban blocked her from getting an education; she’s now a sophomore at MIT. A man learned math on Khan Academy while he was incarcerated for over a decade; he attended Stanford after his release. Those stories are real, and they’re worth celebrating. But they’re the exception, not the model to plan around. For the vast majority of students, the technology has not been the thing that has made the difference on its own. The difference is due to a school leader who commits, a teacher who’s supported with training and empowered to use it, and a system built with enough care that the technology actually gets used the way it’s meant to be used. Give the technology too much of the credit, and you’ll go looking for the wrong lever the next time you try to replicate the result. Give the humans their due, and you’ll know exactly the system that has to be built, and how much work that really takes, to help more students succeed.
