I recently interviewed So. This transcript has been condensed.
Q. This book begins with a confession. You noticed your thinking was changing due to AI. Can you describe what it felt like?
A. I was writing an email to a colleague. I wrote up a partial sentence, and then pasted it into ChatGPT and said, “Can you help me finish this professionally?” And I remember sitting back in my chair and feeling the weight of having essentially removed my own mind and thought process from, you know, the act of work. It was this very uncomfortable balance between, on one hand, being more efficient, and the sense that it was no longer reflecting my own brain, my own originality, and my contribution.
Q. In the book, you describe how naturally AI dependency can take hold. How does it usually happen?
A. So the first stage of the cycle typically starts out rather innocently, where someone begins to use AI. Maybe it’s out of fascination. Maybe it’s because of pressure from their employer. But when you start to use AI, what happens is that you lose touch with the act that typically helps maintain and build skill sets. This notion of “use it or lose it” applies not just to our muscles, but also to many of our cognitive functions.
And when we lose that skill, it then creates this stage where we become uncomfortable doing the activity without the AI, simply because we become accustomed to it. It strikes us as the easier path. Then we reach for AI even more, and the cycle tends to spiral with each pass. It’s not really about any one individual act. It’s really the accumulation of these actions.
Q. The book highlights ways that AI chatbots can be provoked into stimulating our minds rather than dulling them. What I found interesting, and what you point out in the book, is that it isn’t all that difficult. Designers can ask a chatbot to offer hints rather than answers, or introduce a short waiting period between responses. The problem is, the result of these interventions look like a bad user experience, and people also rate the entire program more poorly. Yet such designs create better thinking and better outcomes. Do you think that interfaces that promote thinking can still win?
A. There’s this duality, where many of the things that make us better thinkers are things that we tend not to instinctively like, things like friction or cases where AI acts as a thought partner rather than an answering machine. A lot of AI systems provide these frictionless solutions because they’re trying to drive engagement.
The reason why I still think these systems can win out, even with these frictions, is that once people begin to care about what’s going on in their minds, then it really behooves us to create systems that keep us sharp and that make us think, rather than just giving us an outlet to outsource our thinking.
Q. So you think it’s just a matter of time, as people adjust to the technology, until they see for themselves why the friction was important?
A. Yes, exactly, both from an economic perspective, because you need these things to do important tasks well, but also just from a personal perspective. When you stop humans from engaging in the thought process by making it frictionless, you lose a lot of the things that make us cultivate enduring skills, and I think people are beginning to realize what’s at stake.
Q. In the beginning of the book, you define AI reshaping your mind as feeling a sort of splinter, but by the end, you define it as a loss. Can you talk about what caused that shift in thinking?
A. I used this reference to a movie that I watched growing up, “The Matrix,” where Morpheus says to Neo, you feel it “like a splinter in your mind,” referring to this nagging sensation that something is disrupting your thoughts. I originally thought of it as kind of a foreign presence, like an AI directly operating inside my mind.
But what I’ve come to realize is a lot of the discomfort was not from something like a foreign entity intruding into my mind, but rather the absence of my mind working the way that it used to. With AI, I was able to achieve things, maybe a presentation or a document, without actually having gone through that thought process to produce it.
This is what I ultimately recognized as the sense of discomfort that I think many people feel. It’s almost like a disconnect between what you’re producing and what your mind is working on, and I think that really was a big awakening to me. The risk of these AI systems is that they’re so efficient and so persuasive that they tempt us to essentially remove our minds from day-to-day operations in the normal act of life. That’s a very scary proposition to me, both as an educator and as a parent.
Q. What surprised you while researching this book?
A. What really surprised me was how many people are already feeling that their minds are operating differently in the short amount of time since AI came into the mainstream.
People see it not only in themselves but in their children. I didn’t write about this in the book, but I’ve had so many parents come to me as I’ve presented at various audiences, talking about how they have discovered that their children were outsourcing their brains to AI, and they just didn’t realize fast enough.
Essentially, the kids had learned to rely on AI, to become dependent on it and not be able to think without it. This is a technology that hit the mainstream in very late 2022.
People have talked to me about how they lost touch with the skills they’re meant to be cultivating. I had one terrifying conversation with a medical student who mentioned to me that it had just become their default, any time they were being asked to provide a diagnosis, to ask an AI bot to do it for them. They were basically putting people’s health at risk because they were so tempted by AI too early on. That really jarred me.
Q. What do you think should be the main takeaway for all readers who want to use AI as a tool and not a crutch?
A. If I were to give one critical takeaway, it would be to really think about who you want to be cognitively, five years, 10 years from now. I talk a lot in the book about this notion of “signature thinking skills.” These are the core competencies that I want to cultivate for myself. So as a professor, I think of these skills as being able to distill information from literature to convey what’s important and realistic to people. For a doctor, it may be their diagnostic eye. For a lawyer, it may be rhetoric and discourse.
Once you figure out the things that you really want to value and cultivate in yourself, you can work backward to figure out how you should be using AI. When I’ve talked to people about the book, they’ve realized that they were relying on AI to do the tasks that are often the basis of forming those “signature thinking skills.”
If, as a professor, I have AI do the work that helps cultivate my ability to see the bigger picture and distill research themes, well, then basically I’m eroding the thing that I want to be known for, that I think makes me valuable. I think that’s a really bad use.
Q. Was there anything else that you wanted to cover or reiterate before we end the interview?
A. So many people these days are just really focused on the productivity benefits of AI and what it can do for us. I think what people really should be thinking about is what AI is doing to us in the immediate term.
AI is changing who we are as thinkers, right now, in our day-to-day lives; in classrooms; in households; in boardrooms across the globe.
Rebecca Spiess can be reached at rebecca.spiess@globe.com. Follow her @rebeccaspiessl.