Craig Vezina Conversation with Anant Agarwal, May 13, 2020

Anant Agarwal: In early March the world went from three percent learning online to 100 percent learning and teaching online. As universities closed and sent their students home everybody is now learning and working online from companies as well. So that’s been a really big challenge as people have been trying to adapt to this new normal that I think is going to stay in our lives for a lot longer. And therein also lies opportunity as more and more people have gone online they have discovered that hey, this is quite good. Many professors, virtually all professors have begun to teach remotely, teach online and many of the that were diehard no, it has to be in person are now seeing some of the merits of online teaching and learning and many of them are thinking, you know, this is not half bad. In fact, the students are much more comfortable engaging and laughing and discussing with smilies online when they can be asking these questions anonymously. And we’ve certainly seen a massive mushrooming of interest in edX so when you look at the period during COVID versus the period before COVID we’ve seen edX increase enrollments in edX courses ranging from how to be happy to public health to computer science and data science.

So 10X increase in enrollments by learners all over the world and so people are realizing that this is a great opportunity for them to learn something new as well. I think as universities we should seize this opportunity and move toward blended learning. Blended learning gives you this continuity as an educational institution and in many respects it is better than in person learning. So when you combine the two you get the best of both worlds. So this is the moment where everybody has had a chance to look at it and rather than rubber banding back to the old normal, let us create the new normal in a simpler and better world. And the same thing with companies. For companies a lot of it was in person training and in many companies they’re just not focused on lifelong learning for employees as the future of work really requires everybody to upskill and stands for us to embrace online learning and blended learning and just create a whole new and better world for ourselves.

We’re using AI on the edX platform in a number of ways. AI is used in learner support. When learners call in and they have an issue AI technologies are able to help learners very quickly based on previous requests. Even within a course we have hinting frameworks where learners are able to get hints, for example, that are personalized based on some of the questions they ask or some of the doubts that they have. We also have many courses, for example, we have a course at edX from Harvard on super-earths where a machine learning engine uses AI techniques to point the learner to the next topic they should learn or speed up the learner or slow the learner down. So I think we will see more and more forms of personalized learning come in over time, but I think we have to trade them off against the humanity part of it which is we as humans like to move around in groups and personalized learning is beneficial in a number of ways. It let’s me go at my own speed, for example.

But at the same time we humans have a herd mentality. We like to do things together in groups and if you like to do things in groups and cohorts then everybody has to be going at the same pace which is completely opposite to personalized learning. So I think what we will find is interesting melds of the two where purely personalized learning is not the approach nor is completely synchronized cohort based teaching but maybe some topics are personalized and other topics were working together in a group at the same pace. I think some meld of the two is really the right approach I think. And so I think there is not so good online teaching which is giving your one hour online lecture, but on the other hand there are many, many much more excellent ways of teaching online. One approach is active learning where you interweave short videos and interactive exercises so that students get information in small chunks – five, six, seven minute videos and then that is followed by interactive exercises or discussions when they apply what they just learned.

This is called active learning. And cognitive science from as early as 1972 in a very famous paper by Craik and Lockhart showed that learning outcomes are simply better with active learning. And now with real technology we can apply those learnings from neuroscience. We can apply those learnings to online learning dramatically increasing learning outcomes for the future. That is true online teaching and learning, not sitting in front of a Zoom camera for an hour. The world is advancing so tremendously in this digital age in large part because of the availability of data and that ability to use analytics and data science to process the data to continue to improve outcomes. Similarly on platforms like edX we are gathering what I call the big data of learning. We are capturing every single mouse click and using that we create streams and streams of data which we can analyze to understand how people learn, to understand what works and what doesn’t work and then keep improving learning outcomes either in real time or in your real time. So I’ll just give you one concrete example.

So in 2012 Philip Gould was a researcher at edX and he analyzed five million video watching sessions and based on that he wrote a very famous blog post about six minute videos being optimal. And that involved looking at large amounts of data sets, looking at learner engagement and measuring how learners were interacting and concluding that six minute videos are optimal. And so today everybody around the world quotes that hey, six minute videos are optimal. So that result came from edX where a data scientist sat down and looked at large amounts of data and analyzed it to suggest a good approach that would deal with outcomes. That’s just one example. We also have professors conducting A/B tests on the edX platform, the platform naturally supports A/B testing where you can show students randomly two approaches. You have an A group and a B group and then you give them a common test at the end and you see who did better. And so this way faculty can be using educational engineering to continue to improve learning outcomes using A/B testing on our platform.

And the platform supports A/B testing in a natural way and so all of these techniques use data to continually improve outcomes to create a culture of continuous improvement. So edX is a nonprofit. We were founded by Harvard and MIT and when we started it took us exactly three seconds to determine that we should be a nonprofit. We knew at that time when we started edX that applying digital technology to education was going to completely revolutionize education and we felt that this technology needed to be in the hands of the nonprofit so that we would continually be doing things in a way that was in the best interest of the world, best interest of learners all over the world. And so a number of decisions over the years have come from this foundational factor we are a nonprofit. So as an example we made the edX platform open source. There are a number of for profit vendors in the MOOC space, in the online space and edX is the only nonprofit in the space. And guess what? Ours are the only platform that is open source. All of the others for profits have proprietary platforms. Why would you give away your proprietary algorithm? You would never see Google give away its search algorithm. You would never see Coke give away its Coke formula. These are proprietary.

Now as a nonprofit we’ve given away our secret sauce. We’ve given away our platform code for free and today edX plus its open source platforms have become the world’s largest learning platform. There are 30 million learners on edX and there are 60 million additional learners on these MOOC platforms all over the world with a total of approaching 100 million learners today. And so we believe that as a nonprofit you think and behave differently and education being a human right it is very important that a nonprofit be evolving learning in a way that is ultimately good for the learner as opposed to being good for the bottom line. So we believe really strongly that education and particularly disruptive innovation in education should be coming from a nonprofit where digital technology applied to learning can completely transform the way we learn either in the corporate world or in schools or in colleges. And this has happened in many industries. So just as an example in the hospitality industry we’ve all seen how Airbnb and the likes of TripAdvisor are a platform company and they connect people that want their service to people that have a supply.

But Airbnb or TripAdvisor don’t own any hotels. They’re a platform but they’ve completely transformed the world. Similarly in education we view ourselves as a platform company, a nonprofit platform company where we have a platform connecting learners to partners that have teaching to provide. I believe that we’re just scratching the surface of innovation and here’s one example of something very radical that I see happening in the future. Imagine if all credentials became modular. Instead of creating full degrees what if we created modular credentials. On edX we’ve launched a couple of them. One is called Micro Masters and one is called Micro Bachelors. And the masters level it is about 25 percent of a masters degree and the bachelors level is about three to six, nine credits worth is a Micro Bachelors. What you’re seeing happening now is sharing. As universities are now adopting each other’s MicroMasters or their own degree seeking students, you’re seeing universities share content among each other.

You’ve seen examples where a university might launch a masters degree on campus but have their students take 25 percent of the degree from a different institution. And we see this sharing economy happen which is simply better for students. It will enable better quality. It will enable universities to launch degrees faster. It will enable people to share and sort of create this network digital economy and that’s at the heart of so much innovation that is happening around the world. And I believe this modular stacking, sharing, networking – I believe once that comes to universities and education, we will see just a whole step function increase in the quality and the **** that students will get. How does all of this relate to the future of work? There’s been a lot of discussion about the future of work and this was before COVID-19 where people talked about the work completely changing and automation and AI transforming more than 50 percent of the jobs around the world, the newer jobs. And a lot of worry about what’s going to happen to all of us.

And truly the solution to that was learning. People becoming lifelong learners are where gone are the days you learn for four years while you’re 18. Rather you have to become continuous lifelong learners to be continually upskilling for the future work. And so we need to move to a culture of lifelong learning and frankly, once you’re working, for example, you have a family and so on frankly, the only way to learn is online. I don’t see how you can be going back to a campus for a year or two years, spending another $100,000. I just don’t see that in the cards. So I think online learning is an enabler for lifelong learning and lifelong learning and the future of work are really joined at the hip and just extremely important as we need to contemplate a planet that we have to upskill completely with the future of work in AI and frankly, during COVID-19 as the whole world is now getting a taste of online learning I think this is a good thing for the world in that they’ll be better prepared as they come out of COVID for the future of work.