AI tutors aren't a silver bullet, nor should they be
Personalization in learning is good, but it isn't education's biggest failing
The debate around AI tutors wouldn’t be so aggressive if the claims and expectations were rooted in appropriate use-cases. Ideas from technologists trying to show how all of education will improve for everyone are often talking about content challenges. Their central argument is that more time getting personalized content focused on specific challenges in a subject will reverse the dwindling performance metrics in education.
This bravado inherently creates an opposing party. Those labelled as “skeptics” of AI in education are made to seem like luddites opposing the calculator. Except the calculator was never sold as the instructional panacea the way these AI tutors are. Those who are skeptical of AI’s benefits are often concerned about the relationship that the student forms with a teacher or tutor. It’s not enough that the content be addressed, but also how a good relationship with an instructor helps motivate and engage them with learning.
As an instructor, it’s unfortunate that asking if a technology can really help my students gets me labelled as a luddite. While examples of AI helping a student who is struggling with a large division are certainly inspiring, that single moment in a learner's life is not their entire educational journey. These use cases are derivative of what search can do today. Heck, even Google now has this as a first option in the Google app on Android.
Josh Brake, John Warner, and Dan Meyer are the three best voices I've seen on why this isn’t enough.
They gave me words I didn’t have for that uncomfortable feeling I would get watching demos of AI tutors. Now I know that feeling is due to the missing relationship building with an AI tutor. Relationship building is critical for a student's motivation and engagement—two essential factors for their educational successes. Personalized content in a subject is not the hallmark of a good relationship, but rather only one aspect of better education. Moreover, boring examples aren't the reason why so many students become disenchanted with the educational system. It’s dubious that AI can be effective at helping students overcome struggles in the long-term if the core promise is so narrow.
Essentially, our conversations around AI’s potential in education should always be rooted in who we’re trying to help and how we’re trying to help them. There’s a disturbing emptiness to blanket statements like “everyone would be better off with a tutor, ergo AI tutors are gonna save education.” It demonstrates a lack of empathy on the part of those who believe it. It makes it seem like they don’t understand what it means for someone to learn and the effort that goes into teaching.
It’s also important to highlight that this post isn’t a takedown of AI in education. In fact, I’ve used it in several ways that have been valuable to my instruction. Instead, I’m raising a flag that insufficient AI isn’t anywhere near the top of reasons why education has a problem. Yet instructors are once again being bulldozed in the name of a technology that is both unproven and unemotional. They will once again be misunderstood by everyone too dazzled by a tech demo that has no teeth, and the last thing we need more of in education right now is yet another reason for a lack of understanding.
