In 鈥淪traight Talk with Rick and Jal,鈥� Harvard University鈥檚 Jal Mehta and I examine the reforms and enthusiasms that permeate education. In a field full of buzzwords, our goal is simple: Tell the truth, in plain English, about what鈥檚 being proposed and what it means for students, teachers, and parents. We may be wrong and we will frequently disagree, but we鈥檒l try to be candid and ensure that you don鈥檛 need a Ph.D. in eduspeak to understand us. Today鈥檚 topic is deeper learning in the age of AI.
鈥搁颈肠办
Rick: Jal, not long ago, we discussed the joys of coaching. Our conversation got me thinking about a point you made several years ago in your book . You noted that the best teaching and learning is frequently found not in core academic subjects but in the arts and extracurriculars. That鈥檚 where you see models of intensive coaching and mentoring.
You wrote about how the intrinsic passion that draws students to sports or the arts can empower 糖心动漫vlog to set and enforce rigorous expectations. At a moment when schools are grappling with chronic absenteeism, student misconduct, and grade inflation, there鈥檚 a crying need for both intensive coaching and expectation-setting. Given that, I鈥檓 curious what lessons we might draw from your work.
For instance, one perennial selling point for classroom technology is that it鈥檒l allow teachers to offload a variety of rote tasks, giving them more time to mentor and coach. We all know the mantra: By freeing teachers up from busywork, digital tools will give them more time to be hands-on with students. In practice, though, it rarely seems to work that way. Instead, we see too many elementary classrooms where students navigate their iPads in silence or high schools where teachers struggle with balky tech while students tune out.
That sure doesn鈥檛 add up to hands-on coaching. So, what鈥檒l it take to make intensive coaching鈥攖he kind we see with competitive debate or soccer teams鈥攁 classroom reality? Are there things that administrators can do? Are there lessons that teacher trainers should be taking? I鈥檓 curious to hear your two cents.
Jal: A topic close to my heart! Yes, in our research for the book, Sarah Fine and I found that in the arts and extracurriculars, students were often engaged in the deepest learning. What were some of the features that enabled that? In these spaces, students learn by doing, often from older peers. Moreover, there is a clear sense of purpose and an authentic audience. Students we spoke to said the nature of the community was different in these spaces鈥攊t often felt more like a family than a class. And, as you point out, everyone is choosing to be there, which creates a mission-driven community and accountability for doing what you鈥檝e signed up to do.
We certainly could do more to foster these spaces before the closing bell. I worry that we are enthralled with 颈苍苍辞惫补迟颈辞苍鈥�adding bells and whistles onto our existing structure鈥攁nd too timid to engage in transformation鈥攃hanging those structures in ways that would support deep teaching and learning. There are some simple changes we could make to existing classes鈥攍ess breadth, more depth; less teacher talk, more active learning; fewer worksheets, more projects鈥攖hat would start to unleash the student passion we see in extracurriculars and the arts. Once we make the student the producer of an authentic piece of quality work, then everything starts flowing in the right direction: The teacher鈥檚 role shifts from instructor to coach, peers move from competitors to teammates, and mistakes become a normal part of learning rather than something to be feared.
What鈥檚 the role of AI in this world? Well, if you visit secondary schools that foster deeper learning, you will see that they resemble modern workplaces, with students often working together on long-term projects. In these settings, AI can be used the same way you use it in a workplace. Sometimes, it enables you to do something faster or more efficiently, sometimes, it is useful for brainstorming, and sometimes it is useless. The role of the teacher as coach, then, is to help students make judgments about whether, when, and under what conditions AI can help.
What do you think, Rick? Do you see constructive possibilities for AI? Or should we be resisting it, my 鈥渙ld school鈥� friend?
Rick: Maybe you see it differently, but I find the enthusiastic talk about the promise of AI mostly skipping past the points you just raised about deeper learning. While I鈥檓 dazzled by so many things that AI can do, it鈥檚 not at all clear to me that it鈥檒l necessarily help students learn by doing, learn from older peers, possess a clear sense of purpose, face a meaningful audience, or enjoy a supportive classroom community. Those, of course, are precisely the things you just tabbed as the keys to making deeper learning more than a slogan.
In fact, when it comes to a meaningful audience or supportive classroom community, AI-infused instruction could tug schools in the wrong direction. After all, if students spend even more time working on one-to-one devices or getting chatbot tutoring in their earbuds, it鈥檒l aggravate isolation and further unravel any sense of community. Even where those approaches show a short-term bump on performance, they may come at a price that is not immediately evident.
At the same time, it鈥檚 easy to see how AI could help promote deeper learning as you describe it. I鈥檝e often noted how the ability of athletes to screen game tape on an iPad has allowed coaches to spend more time mentoring players and explaining technique and a lot less energy drawing on chalkboards, assembling playbooks, or running projectors. The same obviously applies to the instructional core, a point which tech aficionados have made time and again.
The problem is that we鈥檝e got a dismal record implementing technology in schools. Somewhere along the way, the transmission line seems to fizzle out. With smartboards, teachers lamented that they spent more time fighting the equipment than delivering dazzling lessons. With one-to-one devices, classes full of elementary students ended up reading digital textbooks or completing digital worksheets while teachers strolled the aisles. The hypothesized returns have been hard to find.
When it comes to AI and coaching, it鈥檚 easy for me to imagine the plot getting lost. Setting aside student-facing AI and returning to the claims for teacher-facing AI, AI-generated lesson plans, grading rubrics, exercises, quizzes, IEPs, parental communications, and feedback could give teachers more time to engage with students. I get it. But I can鈥檛 help wondering whether all this 鈥渇rictionless鈥� preparation may unwind the rhythms and relationships that undergird coaching, especially for new teachers who won鈥檛 have years of institutional memory to draw upon.
I guess I鈥檓 not sure what to make of this. Am I wildly off-base here? If not, any advice on how schools can use the introduction of AI to promote the kind of hands-on mentoring that ed-tech enthusiasts have long promised but too rarely delivered?
Jal: I think you are right to be concerned. Learning, whether it is for teachers or students, is about the learner forcing their minds, hands, and hearts down new pathways. It is a project of expanding yourself鈥攍earning to do things that you couldn鈥檛 have done, or even imagined, before. And there are no shortcuts.
Speaking of coaching, as I watch my middle school basketball players very gradually expand their skills鈥攅.g., learning how to dribble with their off hand, or making a step-through move to the basket鈥攊t is very clear how much training and repetition it takes to do even fairly simple things. So, if you are a very experienced teacher and you鈥檝e designed thousands of lessons, perhaps you can use AI to more quickly make a chart or slide for class. But for newer teachers who haven鈥檛 yet put in their 10,000 hours, using AI to design lessons might rob them of the needed learning that comes from making it yourself.
I鈥檇 be more excited on the teacher side if we used it to create new pedagogic possibilities rather than merely automating what we are already doing. Recently, my brother-in-law, Ben Shiller, had AI write a lengthy children鈥檚 story for my son for Christmas. It was an entirely personalized story that featured our family taking a trip to my 6-year-old鈥檚 favorite show. Ben still had to write a lot of prompts and do some editing, but it is an example of something that would have been infeasible without AI. I鈥檓 excited to see more of these kinds of innovations, drawing on a new technology that creates a wide variety of possibilities, rather than substituting AI in situations where humans still need to do the learning.
This conversation has been edited for length and clarity.
