This week in MIT Massive, we had an activity where we were tasked with designing a school which takes advantage of personalized learning (where in this case, personalized learning means software like Knewton or Khan Academy). All the groups in our class assumed that such software can save time used to the teach basic content, freeing up time to participate in deeper learning experiences (e.g., project based learning). None of us designed a school where the personalized learning was both the means and the end, where the core student experience was solely interacting with their computer. Given that we seemed to agree that computers weren’t better at teaching everything, Justin asked if we see a limit to the role that technology can play in the classroom moving into the future. I have some thoughts on the topic that I’ll detail below, but to give you fair warning, my answer to this question is probably more of an excuse to wax philosophical than it is an attempt to reveal any insight about the role of technology in education.
The short version of my answer is that I don’t think there is any limit to the role technology can play in the classroom. In the future, I think it’s possible that a student could get a better education wearing a VR headset connected to a consumer computer than they could get in the best classrooms in the world today. And the version of “better education” that I’m claiming is all inclusive: they would have stronger content knowledge, better critical thinking skills, be more motivated to continue learning, have a stronger internal locus of control, and score higher on any other metric you can come up with. I believe this for primarily two reasons: 1. human consciousness does not transcend any physical laws and 2. technology continues to march forward at an unrelenting rate.
For me, the first point is almost self-evident. While consciousness is incredibly complicated, and the current framework for understanding it is beyond me, I do know that consciousness arises from something physical: the connection of neurons in our brain. While it’s true that there are many, many neurons in a human brain (on the order of 100 billion), there are also billions of transistors in the CPU of the laptop I’m writing this post on. From a physics perspective, I don’t see why we couldn’t connect a bunch of computers together and have consciousness arise: physically the infrastructure is there, we just haven’t made the right connections yet. A more visceral manifestation of this idea is that if you walked away from the movie Her thinking that [spoiler alert] Theodore’s love for Samantha (the AI) was at least as real as his love of Catherine (his ex-wife) had been, you were convinced that there is nothing magic about humans that gives us a monopoly on consciousness.
The second point is also fairly self-evident in current society. Moore’s law, the doubling of computational power every two or so years, is a well known phenomenon. A popular statement of the the effect of Moore’s law is that the computer in your pocket today is more powerful (by many orders of magnitude) than the computers used to land a man on the moon. While Moore’s law has been slowing down in the past few years, it doesn’t change the fact that computers are still becoming more powerful, and we can’t imagine a limit to what they’d be able to do (or perhaps more accurately, the limit is effectively infinity compared to what they can do today). One of my personal favorite examples of how far computers have come is AlphaGo, the computer program which was able to beat one of the world’s best Go players earlier this year, a task which wasn’t expected to happen until 2025 by most experts.
If these two points sound like they’re adding up to the singularity, the idea that when computer intelligence surpasses human intelligence it will lead to an infinitely accelerating wave of technological progress, I guess it’s because in essence that is what I’m arguing. If we accept these two points, eventually we’ll have computers that can pass the Turing test which are cheaper than a smartphone is today. Every student could have a personal AI tutor with the capacity for thought of a human and instant access to essentially all of human knowledge. They could interact with this tutor in a virtual reality with a level of resolution indistinguishable from the real world. They could also work on projects in teams where their teammates are AI teachable agents, operating to provide their student with an ideal learning experience (both in terms of learning content and non-cognitive skills). I would not be surprised if this vision is frightening to many, and if that’s the case for you I would remind you that in this post-singularity scenario, interacting with an AI is just as rich and complex as interacting with another real live human being.
Of course, even if this future is possible (or inevitable), it doesn’t do much to inform how technology should be used in education today. Any skill that a computer can teach more effectively than a human can also be performed more effectively by a computer than a human, meaning that skill is not useful for a human to have in order to be a productive member of society. In this future, there would be no need for humans to do work and we could live idle lives full of whatever leisurely pursuits we found most interesting. So while there’s no limit to the role technology could play in the post-singularity classroom of the future, that lack of a limit is irrelevant to the classroom of today.