The Limits of a Mechanical Sparring Partner
Getting personalized help with your work is only a small part of what makes a teacher great
Educators wear a lot of different hats: teacher, instructor, lecturer, coach, colleague, mentor, and friend are just a few in a long list. This summer I’ve been thinking about how the different roles of my job as an educator require different skill sets and approaches. How you should respond in a particular situation as an instructor is not necessarily the way that you should respond as a mentor or coach. And while this delineation of roles is valuable in and of itself, I think the growing specter of AI in the form of large language models like GPT-4 is going to make thinking about these separate roles even more important.
Education is a human activity
Today I want to spend a little time sketching out some thoughts about how the different aspects of education help me to see the humanity in it. In that process, I want to reflect on how AI might be thoughtfully integrated into some of these roles, with a focus on using it as a ladder and not a crutch. And for those of you who are going “I’m not a teacher, peace out!”—hang with me. I hope that you’ll realize that all of us are educators in some context. Even if you aren’t working in the context of a school, these principles can still help you to be a better colleague, parent, and friend.
Education is a multifaceted enterprise
As AI continues to permeate educational spaces, it is worth thinking about what our jobs as educators are really about. Teaching students new knowledge might be the most significant aspect of our work as educators. But I’m not so sure. There are other parts of the job that I think are just as meaningful as teaching students new knowledge. Like helping them to discover their interests and explore who they are as they enter the season of emerging adulthood. Or asking insightful questions that might help them to discern what’s important in their lives and where they want to focus their time and attention. Or just being someone who cares enough to get to know their name and give them your time and care. Recognizing the different facets of education will be increasingly important as we see AI continue to become more and more embedded in education at all levels this fall.
On the AI in education beat, Harvard recently announced that this fall they will experiment with an AI tutor in their popular introductory computer science course, CS 50. The goal of the AI tutor, built on a customized large language model (think custom ChatGPT) is to help nudge students in the right direction toward solutions as they work through questions. Professor David J. Malan shares:
Our own hope is that, through AI, we can eventually approximate a 1:1 teacher:student ratio for every student in CS50, as by providing them with software-based tools that, 24/7, can support their learning at a pace and in a style that works best for them individually.
While this might start with Harvard, it doesn’t stop there. Not even close. Chatbot U is about to pop up all over the place this fall as
insightfully pointed out this week on his Substack. These chatbot interfaces are not only being deployed within the walls of Harvard but also are a central piece of Khan Academy’s strategic goals as demonstrated by their recent partnership with OpenAI. Khan Academy’s founder and CEO, Sal Khan, makes the case for integrating AI into education in his 2023 TED talk.Pretty soon AI tutoring will be everywhere. And given the cultural hype and money behind these tools it’s hard to imagine that they won’t soon be seeping into your educational world sooner rather than later.
Even good intentions can lead to negative impacts
Setting aside the many ethical questions related to AI like the biases embedded within them (both intentional and unintentional), the data privacy and intellectual property concerns about the data they are being trained on, and the glitches that cause these models to spit out factually false information, I can understand the appeal. Technology can be a force for democratization. Private 1-on-1 tutoring can be extremely valuable for those who are fortunate to get it. Using technology, AI or not, to take these experiences which have historically been available only to a privileged few and deploying them at a broader, more accessible scale is a worthwhile pursuit.
But at the end of the day, we need to sit down and think carefully about what the move to replace humans with machines means. What are we saying about the value of an actual human interacting with you? Do these machines really “approximate a 1:1 teacher:student ratio” as Prof. Malan suggested? In fairness to him, I take his statement to be exploring whether using AI might help to free up more human time to do the work that is uniquely human. But what exactly is that? Is the problem one that is best solved with bots or by hiring more student TAs?
There is a danger in seeing educational problems as ones that can be solved with technology. This perspective colors our view and makes us prone to forget that education is not just about uploading data into the human brain. At its root, education is a human activity. It takes place within a relationship. Robots know if your code doesn’t compile and how to give you a hint to fix it. But robots have no way of knowing that the reason you’re not doing well is because you’re dealing with a conflict with your roommate, an argument with your parents, or the loss of a family member. Or that you’re just sleep deprived. We’re not just brains on a stick and the reasons that we aren’t doing our best in class are never solely intellectual.
Time will tell. It’s possible that these AI tools will help to approximate and improve some aspects of education. I can imagine that an AI’s ability to tirelessly ask questions to help students explore areas where their understanding is weak could be a valuable tool. But teaching is much more than tirelessly asking questions.
So what?
I’m not anti-technology. I’m an engineer after all. I also use technology in my teaching, for example, to facilitate teaching in a flipped classroom format. But the crux of the issue is why. In my case, the flipped format allows me to spend more time solving problems with students in class, the part of their learning that can be most frustrating if done in isolation. By the same token, I’m not opposed to using AI in my teaching. Even tools like ChatGPT in their current form can help you to learn and reflect on your work. But how I choose to use AI must be dictated in the same way as any other technology: does it help humans learn?
Being able to approximate one aspect of a teacher is not the same thing as being a teacher in the same way that putting words together in a way that is syntactically correct is not the same as conveying a thought. Only humans make sentences.
The only way ChatGPT wins is if we allow our pleasures to be defined down to what the AI can do, rather than to continue to revel in what humans make.
Maybe AI tools can ask good questions to serve as mechanical sparring partners. But for the other aspects of the teacher-student interactions like mentor, coach, and friend? I’ll take a human.
The Book Nook
I’m currently in the middle of Culture Making by Andy Crouch. I’ve read a few of Andy’s other books but finally am getting back to one of his earliest and best-known ones. While I’m still not quite halfway through yet, I’m really enjoying the way that Andy draws out the various aspects of what culture is and how we are creating it.
A few quotes that have resonated with me so far:
There is no such thing as “the Culture,” and any attempt to talk about “the Culture,” especially in terms of “transforming the Culture,” is misled and misleading. Real culture making, not to mention cultural transformation, begins with a decision about which cultural world—or, better, worlds—we will attempt to make something of.
The biggest cultural mistake we can indulge in is to yearn for technological “solutions” to our deepest cultural “problems.”
The first responsibility of culture makers is not to make something new but to become fluent in the cultural tradition to which we are responsible. Before we can be culture makers, we must be culture keepers.
The most important things in our life are learned by trust, not by deduction from experiment.
The Professor Is In
This week I’m busy putting the finishing touches on a proposal that is due at the end of the week. It always feels like crunch time at the end of the game but I do find that the pressure of a deadline brings out the best in me. Wish me luck as I work on cranking out and clearly communicating my ideas!
Leisure Line
This week at Costco I finally managed to talk Mrs. Absent-minded Professor into the fancy toaster oven that I’ve wanted for several years now. Now I’m enjoying the little luxuries in life: like toasting my bread for breakfast in four short minutes while I fry my eggs.
Still Life
As we flew out of JFK last week we had to take a pit stop on the tarmac to check a maintenance light during our quick 1-hour taxi before taking off (luckily all was well). Glad that this long line of planes waiting was gracious enough to allow us to cut back in at the front.
"The first responsibility of culture makers is not to make something new but to become fluent in the cultural tradition to which we are responsible. Before we can be culture makers, we must be culture keepers."
It strikes me that this approach is analogous to the LitReview part of scholarly research and articles. Is there any part of human endeavor where at least some familiarity with the tradition is not at least somewhat useful/relevant?
Years ago, I had the good fortune to work with a supervisor who was incredibly knowledgeable about his (non-academic) work and who had a caring heart. He draped those, however, in a continual series of quips which ranged from cynical to caustic to delightfully profane -- almost always humorous -- as he tried to lead the office through another day.
One of his quips I was intrigued by was "What goes around, comes around". I used to think to myself -- this guy (younger then than I am now, lol) knows *history* -- the tradition, context, landscape, the successes, the failures, the skeletons in the closets, etc. Over his 30+ years in the business and nearing retirement, he had "seen it all" and had an expansive appreciation of what strategies and techniques had worked and hadn't worked.
Of course, each new idea needs to be re-evaluated in light of contemporary circumstances (which may be very different from past efforts), but I used to think then (and have only grown in appreciation since) that we should have listened to him more closely.
I am interested in critiques which are able to unpack today's useful ideas and follies. At the risk of oversimplifying, I've read pieces about "Airbnb reinventing the hotel"; "Uber and Lyft reinventing the taxi cab or, with their plan for scheduled stops, the city bus"; "Cryptocurrency reinventing the pyramid scheme"; etc. A sign of age, I know, but I hold fast to my twofold belief that there *isn't* an app for everything and that we should never stop trying.