Will AI transform education? The case against

A previous blog post noted the recent advances in AI that are being noted across many sectors, including in education. It’s not surprising that some people are predicting a near-term transformation of education as AI spreads. This post explores why those predictions are probably wrong. Next week’s post will explore why, maybe, they could be right.

The simplest argument for why AI will not transform education is this:

We have at least 100 years of people predicting that a new technology will transform education. To this point, those predictions have a perfect track record of being 100% wrong.

If you’re inclined to view this argument in video form, here it is in just over seven minutes. As the video points out, some of the technologies that have been predicted to revolutionize education include:

• Movies

• Radio

• Television

• Computers

• Videodiscs

• Smartphones

• Tablets

• And, more broadly, the Internet.

Among the themes of many of these technologies has been that they allowed “the” expert in each subject area to reach millions of students, giving access to the “best” teachers regardless of where they live. The technology advocates largely if not entirely overlooked the fact that the best teachers see themselves as teaching students and not teaching content, which means prioritizing many elements of the students over the nuances of the material.

To be clear, I’m not suggesting that online learning has had no effect on students. Online learning is best understood as an instructional modality, not a technology; a modality that allows for the flourishing of new ways to reach and engage students by breaking down barriers of time and place. As many sessions at DLAC will soon demonstrate, online and hybrid programs have positively impacted millions of students. But it’s far too early to say that digital learning has revolutionized education, because the percentage of students who have experienced a transformative online/hybrid class or school remains very small.

Also, it’s important to recognize the difference between enhancements and incremental improvements—which technology has definitely provided—and transformation, which it has not provided.

Given the long history of predictions that the latest technology will transform education, the starting point for anyone who is saying AI will change education is to explain why this technology is fundamentally different than all the technologies that came before. To this point, the case is not compelling. AI advocates have predicted that adaptive learning would transform classrooms for many years, but few predictions have come true, while failures like Knewton are more common. Note also that educators and less-interested observers may not recognize Knewton as the complete failure that it was, for two reasons. First, the hype as Knewton was growing exceeded the attention when Knewton failed. Second, Knewton’s failure took the form of a sale for $17 million, which might not sound like an obvious failure unless you know that investors poured $182 million into the company.

In summary, the case against AI being likely to transform K-12 education is that a long line of “transformational” education technologies has failed over the last century. That evidence alone suggests that the most likely outcome is that AI will be yet another over-hyped technology that has, at most, marginal impacts on education.

But maybe this time is different? Next week’s post will explore that argument.