AI Education: Reshaping Personalized Learning

AI Education: Reshaping Personalized Learning

In AI by Daniel NewmanLeave a Comment

AI Education: Reshaping Personalized Learning

Are teachers the next to be deemed irrelevant by AI and automation? Possibly, if Squirrel AI has anything to do with it. The company is taking China by storm by using AI to provide personalized tutoring for students—primarily to improve their standardized testing scores. And thanks to a new partnership with Carnegie Mellon University, the company may soon be bringing that technology—AI education—to the United States, for better or worse.

First things first: the good news. Squirrel AI isn’t gaining attention for no reason. Two million have already used their AI education tools to their advantage, and some 50,000 current students are using their technology throughout China right now to amazing results. One student, for instance went from scoring a 50 percent to an 85 percent on his math exam within just two years. For him, the tech was the difference between success and failure in a country where huge emphasis is put on standardized test results.

How does it work? Squirrel uses AI education tech to break math concepts into thousands of tiny micro-elements in an effort to pinpoint precisely where a student is missing the boat, and to design a program to help him or her hop back on it. For instance, their middle school math program alone has 10,000 elements—about 10 times as many as the average textbook does. Using a diagnostic test, the company can map knowledge gaps within just 10 questions. Even more shocking: in October 2017, the company ran a study that showed its IT can also help students fill those gaps better than experienced teachers can.

Now: the not-so-good news. First, as I just noted, the success Squirrel AI has seen in preparing Chinese students for standardized testing may lead some countries—including the United States—to begin focusing even more on standardized test results in a time when we all know that they are far from indicative of a child’s actual learning, or ability to learn.

Second, also as noted, if we do begin to focus on standardized testing, we may begin to give our teachers even less credit and credence than we already do, in a time when they provide so much, for so little, for our children. As one expert has already noticed, there is a big difference between personalized education and adaptive education. Squirrel is focused mostly on adaptive learning—finding out where students are struggling and adapting curriculum in such a way that they can learn to score well on it. Personalized education, however, is more of getting to know a student—finding out what makes them tick—encouraging them to pursue those things that use their strengths and personality for the highest purpose. That’s not necessarily what Squirrel AI education does. And it’s not clear yet whether its U.S. lab, which Carnegie Mellon says will provide “personalized education at scale” will either.

Will AI education take over … education? Not likely in the next decade, at least not in the United States. However, it’s imperative that before it starts to infiltrate our classrooms, we create a thoughtful strategy surrounding how to use it in the way that is best for the learning of the child. The way Squirrel AI operates, the learning process for our children becomes a sort of block box scenario. Just like we don’t understand how AI learns, we as parents, and our children as students, aren’t a big enough part of the learning process to really understand how they are learning or why. Obviously, this has potentially harmful implications as new effective algorithms are developed.

Squirrel says it can reduce the amount of unnecessary tutoring by 80 percent. Students at its 2,000 training centers throughout China (yes, 2,000) may agree. My only question, however, is what is lost in those learning hours that have been eliminated. Will students lose the ability to determine on their own where they are struggling? Will they lose the ability to describe to a teacher what they are having difficulty understanding? Will they lose the ability to speak through a problem, rather than simply mysteriously beginning to understand it? Will they lose, even more, socialization and creativity that is disappearing from our society at an alarming rate. After all, in the future we don’t need more people who thing like robots. We need people who can think creatively … critically. It’s not clear how Squirrel AI, or the 60 other companies like it in China, will contribute to that goal.

The original version of this article was first published on Future of Work.