Most of the online courses that currently exist, including the MOOCs that have received so much attention, largely take advantage of the recent developments I’ve just described—broadcasting educational words, sounds, and images at a scale and cost without precedent. We should not diminish that accomplishment. It represents an amazing step forward, in and of itself.
Yet what most of those courses haven’t managed is using information technology to improve the quality of education—something that hasn’t been done, arguably, since the invention of the written word.
But they will, soon, for two reasons. First, because modern information technology allows for unprecedented interactivity and inter-personal connection, the formation of communities of learners on a global scale. Second, because the ability of computers to process information means that we can, for the first time, replicate and improve upon fundamental processes of learning.
Think again about our supposedly ideal model: master and student, sitting together, talking. It’s not as perfect as it seems. Even the smartest person in the world knows far less than what he doesn’t know. And the implicit process of diagnosis in the tutorial model, in which the teacher listens to the student, makes judgments about what they’ve learned, and responds accordingly, is bound by essentially human limitations. It is very difficult to see inside another person’s mind, to provoke and inspire their learning in just the right way. We value people who are great at this precisely because they are always so rare.
It is no coincidence that all of people leading the major MOOC providers—Sebastian Thrun and Peter Norvig at Udacity, Andrew Ng at Coursera, Anant Agarwal at EdX—come from or near the academic disciplines of artificial intelligence and machine learning. Or that some of the best designed online courses, whose effectiveness has been proven by rigorous research, come from institutions like Carnegie Mellon and Stanford, based on decades of research into cognitive psychology, neuroscience, human-computer interaction, and artificial intelligence.
The amount of data that will be generated by millions of people engaged in digital learning environments will yield insights that no single educator can obtain. The AI-based cognitive tutors that are already being used in online courses today will only get better over time. This is not science fiction. It is the unevenly distributed future, rapidly approaching the universal present.
Traditional colleges and universities, predictably, are following in the footsteps of their predecessors by badly underestimating the net educational benefits of information technology. These are organizations designed in the 19th century under conditions of information scarcity that simply no longer exist. And they are in a profound state of denial about all of this.
To start, they grossly underestimate how much of the education they currently provide is already wholly replaceable by a simple broadcast model. Every aspect of the standard lower-division lecture course (which happens to be hugely profitable for colleges) can be perfectly replicated online today and distributed at no marginal cost. It’s trivial, and it’s already been done. Any of you can, for example, take MIT’s entire mandatory undergraduate science and math curriculum, exactly as the students in Cambridge, Massachusetts, take it, on edX, for free.
If you can get colleges to admit this, they’ll fall back on assertions that are, unsurprisingly, rooted in the intangible, the ineffable, the unprovable and the “you just have to be here to understand.” A few things to keep in mind about this:
First, whatever the benefits of things like the campus environment and interpersonal interaction with professors may be—and to be clear, they are certainly real—colleges have absolutely no evidence that would meet their own standards of scholarship credibly estimating or quantifying the size of those benefits. None. If you don’t believe me, ask them some time.
Second, those remaining benefits have to be understood in relation to cost. As we all know, the cost of college is rising at an alarming rate. Even if colleges can prove how much value they added on top of what information technology can provide for free, that still leaves open the question of whether that value is worth the prices they charge.
Third, artificial intelligence and machine learning will inevitably continue to improve and steadily eat away at whatever plausible remaining value proposition traditional colleges may have.
All of which means we’re in an exciting time to be discussing technology and the future of higher education— the first time since the age of Socrates and Aristotle when information technology will make education not just more abundant but more effective for people around the world.
This article was adapted from remarks Kevin Carey delivered on April 30, 2014, at the Future Tense event “Hacking the University: Will Tech Fix Higher Education?” Future Tense is a collaboration among Arizona State University, the New America Foundation, and Slate. Future Tense explores the ways emerging technologies affect society, policy, and culture. To read more, visit the Future Tense blog and the Future Tense home page. You can also follow us on Twitter.