Self-Driving Education

Sebastian Thrun is a programmer, robotics developer and computer scientist from Germany. He is CEO and cofounder of Udacity, with David Stavens and Mike Sokolsky.

PALO ALTO --These are exciting times for anyone in higher education. If you believe Thomas Friedman from the New York Times, we are at the cusp of a major revolution. Khan Academy, MOOCs, and many other forms of online learning are leveraging digital technology into the core of the educational experience. They are fundamentally changing the laws of access, costs, and the student learning experience across the globe.

In the field of higher ed, many have asked whether (or when) digital education will replace on-campus education. I wonder the opposite. Cinema never replaced theatre. TV didn't replace radio. I wonder how different digital education will be from classrooms, and where it will lead us.

With any new medium, the full power is only unearthed with experimentation. Today's online education still closely mimics classrooms. No one truly knows what the full potential of this new medium might be.

But let me venture a guess.

Many students learn best by doing. But because classrooms force the same pace on all students, they limit the degree to which students can truly learn through trial and error. Instead, lectures still force many students to follow material passively and in lockstep pace.

But in the digital world, learners may progress at their own pace - just as in video games. I view this as one of the most fundamental characteristics of digital education. It's okay if a student takes a little extra time. Mastery trumps schedule. Some students might reach a mastery level in a day, others in a month. Sal Khan has shown that failing math students can progress to the top of their class - something classroom instruction rarely achieve.

The individualization of learning fundamentally redefines the role of assessment. Digital education affords us to let exams drive student progress, instead of lectures.

In particular, to the extent that we can digitize grading and instructor feedback, students can practice many times over again, with instant feedback on their performance. When an exam is mastered, the student has confidence in her abilities. Think: Angry Birds. Each level in video games is just like a little exam. The addictions gamers derive from the constant duality of challenges and success can be transported into the educational experience.

This "formative" use of exams has often been quoted as great pedagogical tool. Digital education is lifting this idea to entirely new levels. One Udacity student blogged about her addiction as "binge learning."

Digital education is also all about Big Data. Every mouse click and every keystroke can be recorded, enabling the learning environment to make powerful inferences about a student's ability to grasp an idea or acquire a skill.

Companies like Knewton are mining this data to develop personalized learning paths that maximize for learning outcomes. No more industrial one-size-fits-all. Adaptive learning will lead to learning environments in which every learner will succeed to their fullest potential without ever being bored.

But the data will also help educators. Every aspect of an educational product -- no matter how small -- can be analyzed. Smart providers will leverage data to continually weed out weaknesses in their approaches, to maximize for student experience and learning outcomes. And as a result, the quality of education will skyrocket.

Large providers will benefit more from data than small boutique educators. And large providers can also invest more resources into the creation of high quality education. Analogous to the consolidation that took place in the movie industry, I predict the digital education industry will be comprised of a handful of major players. Professors will collaborate with teams of specialists driving high quality products and learning experiences.

Global student engagement will open new ways for peer interaction. This might comprise of small, highly selected teams jointly tacking challenges, or partaking in global discussions where students learn with the best in the world -- would they be in Anchorage, Athens, or Ahmedabad.

A final implication of the decoupling of student pace means that learning will blend much better into the physical lives of busy people. Traditionally, we dedicate a distinct phase of early adulthood to higher education, where we hope to learn a lifetime worth of things we ought to know.

The digital era makes it possible to acquire skills and knowledge on-demand, in small chunks, when we need it. Every free hour can now be used to advance one's skills, using our mobile device. Perhaps we can learn while we are passengers in our self-driving cars.

This makes this new medium much more susceptible to lifelong learners, who wish to continue to educate themselves throughout their career. Young people love to flock to campuses. But once in a job with a family and a mortgage, taking time off and relocating to a residential campus can be challenging.

I venture that digital education will be a game changer. In the US, adults stay in the same jobs for an average of 4.1 years. They assume seven different careers during their lifetime. Digital education will provide education to people throughout their lives, to stay current in an ever-changing business and technology landscape.

Digital education might be our best and only hope to reduce the growing job skills gap. McKinsey estimates that by 2020, there will be 85 million open jobs worldwide for skilled and partially skilled labor. The present system offers no scalable answer to this challenge, and CEOs of major corporations are desperate.

I admit this is all just a guess. The path of innovation often leads to surprises. But I am convinced that this new medium opens up a potential to redefine the basic fabric of higher education. As a committed educator, I am 100 percent dedicated to explore this new medium, to the full benefits of all the students in the world.