In 2006 Salman Khan logged on to YouTube and uploaded a handful of videos he had made to help his cousins with their homework. Three years later Khan’s videos had so many users that he quit his finance job to focus full-time on his free, nonprofit online education project. Bill Gates and Google soon got involved, and the media spread Khan’s story far and wide. Today six million students visit the Khan Academy each month where they watch lectures on subjects ranging from arithmetic to differential equations. Scientific American talked to Khan about the state of American education, his critics and the future of the Khan Academy.

[An excerpt of the edited transcript of the interview follows.]

What do you think is the primary problem with education in the U.S. right now?
People are asking the wrong questions about how education can be improved. Everyone assumes the model of kids being batched together in synchronous cohorts. Credentials, for the most part, are based on seat time versus competency. I think students should be able to learn at their own pace. They should master concepts before moving on. Credentials should be competency based. If you learn something on Khan Academy or at your community college or at Harvard or on the job, you should be able to prove it, and also show that you’ve retained it over a period of time.

Beyond flipping the classroom—watching video lectures at home and doing homework on class—how do you think Khan Academy addresses those issues?
We’ve kind of been glued to the idea of flipping the classroom because, eight minutes into my TED talk, I talk about it. But frankly, we don’t really view ourselves as a flipped learning model. Flip is kind of a neat idea, and I think it is better than giving the lectures in the classroom and doing homework on your own. But we consider ourselves advocating a self-paced, mastery-based model that frees up teacher time and class time to focus on higher-order tasks.

Our team is now pushing 40 people, and the site is changing as we speak. I’m not the only one creating video content anymore, and the interactive platform is getting much more substantive. In the next nine months you should be able to go to Khan Academy site and say, “I’m interested in algebra,” and we’ll give you a state-of-the-art, adaptive test that will pinpoint what you know and what you don’t. Based on that, we will recommend activities we think will be valuable for you. We will try to pace you, to be a virtual tutor.

There have been some great stories about really young kids using Khan Academy to learn calculus when they’re 10. What happens when they’re 12 and doing graduate level mathematics? Are there limits to how well self pacing can work in a traditional classroom setting?
My take is, if a kid’s doing that, that’s awesome. They should keep pushing the envelope. Depending on where a child is, maturitywise, it’s probably a good idea that they stay at their traditional school, with similar age cohorts. But they should have an outlet, several hours a day, especially if they’re off the charts in math or science or something else. Those students who, by age 16, know as much as a normal computer science major—for the summers let them get an internship at Google. Let them start doing some stuff with that knowledge. That only helps society.

Who is this approach better for—those 16-year-old Google interns or kids who are struggling?
I would have assumed, going into this, that it would have been the kids who are classically self motivated. But we are seeing that it’s great kids who are demotivated or checked out, or who have a lot of basic, fundamental weaknesses that, in a traditional class, they’re embarrassed by. There’s no way for them to address it. It’s hard for the teacher to even diagnose those weaknesses—and now, this is kind of a very powerful remediation tool for those students to do it in a very safe way.

You use the phrase “humanize the classroom” a lot. Can you explain what that means?
Most of the classrooms I remember, and a lot of the ones I visit now, don’t look that different. The kids are in a room. The desks are facing the chalkboard—or I guess now it’s a whiteboard or a SMART board. Most of class time is spent with a lecture and kids of taking notes. Every now and then, there is some interaction. But humans— especially young humans—aren’t good at sitting still and passively listening. They want to do things. They want to talk to each other. They want to form human bonds with each other. So when I think of a human classroom, I think of a classroom that’s not 10 percent asking questions of the teacher. It’s 100 percent asking questions of the teacher, and the teacher asking questions of students, and the students talking to each other, working on things together. That, to me, is a far more human experience.

Are you finding that teachers are open to this?
I haven’t had someone stand up and say, “No, no, no. I think actually it should be about the lecture.” The teachers who are using us, we’ve heard nothing but good things about how they feel more empowered. Where I would say where there’s resistance is more of just how people perceive us. A lot of the press narrative has been around the videos. That was minute eight of my TED talk. The next 10 minutes was talking about how we optimized human interaction and allowed students to work at their own pace and armed students and teachers with data and all these other things. Because so many people think that teaching is about lecturing, they think that this guy who’s making videos is trying to replace teachers with a virtual tool. But that’s not it. This is coming from the guy who’s made 3,000 videos: I don’t think the lecture is the most important part of the learning process.

One other criticism that I’ve heard is that your approach isn’t radical enough.
That’s an interesting one. Some of these people call themselves constructivists or constructionists, and for them it’s all about students learning through discovering the process—you shouldn’t kind of feed information to them. And there’s been a debate, especially in math circles and science circles, about what works better—discovery or lectures and problem sets? Our philosophy is that these are not competing interests. These are actually complementary skill sets that students need to pick up. There’s value in being able to do a lot of the classical physics problems. The great physicists of our time learned that way. They had good physics instructors who made them do hard problems.

How far can self-paced learning go? Like what are the institutional obstacles to doing this like public school–wide?
It seems like fastest adopters here are actually some of the leading institutions, either the higher education or the K–12 level. Fast-forward to a world five or 10 years in the future, where the best schools in the country have, to some degree, transitioned to a model where they are not lecturing anymore, where they are a competency-based model. And you have data showing that it’s working. I think it’s very hard for the rest of the system to just say, “We don’t want to be like the best. We just want to do what we’ve been doing.”

So what will Khan Academy look like five years from now?
In five years I would like Khan Academy to be a platform where you can go and say, “I want to learn American history. I want to learn algebra. I want to learn quantum physics.” Pretty much any mainstream topic. And it’ll say, “Okay, great—here’s the diagnostic for you.” You take the diagnostic and then it immediately turns into this kind of active tutor, to pace you and make sure you retain things.

Once you reach a certain knowledge state, the system should be able to tell you, “You know what you’re doing. We think there’s a 95 percent chance that if you were to take the credential test on this right now, you would do very well. It costs $20 to take it at the local testing center. Go do it.”

We’re going to be fully internationalized in five years, so we’re available in every major language. We’re at six million students per month now. In five years I don't know. Fifty million—I’m kind of just pulling that number out of the air—sounds, on one level, aggressive. But it doesn’t sound crazy.

What about a brick-and-mortar school?
Even before Khan Academy existed, I always wanted to start a school. But for the next two to three years the focus needs to be on developing this platform. Then we can create a school: hire some amazing teachers, bring some students in and really start to experiment with the physical experience. Call it what you will—the school of the future.