Cogut Institute for the Humanities

Knowledge as Process

At Brown and across higher ed, AI is the topic of conversation. Some of the questions driving it are practical (How are we using it? How should we use it?) while others are harder to raise (How much of teaching and learning are we entrusting to private technology companies? Why did a technology from Silicon Valley come to be adopted so quickly, with so little discussion of the public good?) And beneath those are existential questions (What becomes of reading, writing, and thinking when a machine can seem to do them for us? What does it mean to be human when the line between human and machine thinking keeps shifting?)

A perspective that rarely enters these conversations, except as a worry about cheating, is that of students themselves. What do they think, and what might we learn if we asked them directly? I had a chance to find out while recording an oral history of the Cogut Institute’s AI and the Humanities Lab, which the historian Holly Case and the computer and data scientist Suresh Venkatasubramanian have directed since 2024. My conversation with two current juniors was the most revealing, informative, and inspiring.

Anika Kotapally studies Chemistry and Comparative Literature, and Chelsea Long studies English and Mathematics. Both joined the lab’s AI and Sci-Fi reading group as first-years, deeply skeptical of AI, and neither has become an enthusiastic user. They wanted to understand AI well enough to argue about it on their own terms, and they don’t claim to speak for anyone but themselves. But as our conversation showed, they think about AI with a seriousness and nuance and are asking the kinds of incisive questions that faculty and administrators should be paying closer attention to.

We spoke on Zoom on October 12, 2026. The conversation has been edited for length and clarity.

Conversation

Shelley Lee: ChatGPT came out in 2022. Before the lab started in 2024, what did you think about AI?

Chelsea Long: My thinking was very abstract and one-dimensional, and uninformed. I knew it was a thing a lot of my classmates in high school used to cheat on papers, in a way I thought was very unwise. I remember people saying, “Oh, I’m going to get it to write up this paper that I have to submit to the International Baccalaureate testing site.” And I thought, “You obviously shouldn’t do that. They’re going to catch you.” So to me, it was just a thing people used to get out of doing work.

Anika Kotapally: I felt very similarly. I had an English teacher in high school who was very concerned about AI, and he was the only person who really talked about it, so many of my thoughts came from that angle. That one-dimensional thinking continued into my first year at Brown. I just thought of AI as a bad thing that I don’t want to use. I knew the general anti-AI talking points: it uses a lot of energy, a lot of water; it raises power costs for people in areas where data centers are being built. I had friends in computer science who knew more about how AI actually worked. But even with them, I was very much, “This is my viewpoint. I don’t really want to believe otherwise.”

Shelley Lee: Was that based on your own experience with it?

Chelsea Long: Overwhelmingly on things I’d heard from other people. I never used AI at any point in high school.

Anika Kotapally: Same. I think the first time I used AI was this year. I did not use AI, point-blank, period, in high school, and I was very firm about that. I was like, “I don’t want to use it. I think it’s bad for my brain.”

Shelley Lee: Were you unusual in that regard?

Chelsea Long: I am unusual. In my close circle of friends, I’m the only one. A lot of my friends are CS majors, so they have to use AI for class. I don’t have a ChatGPT account. My mom and my younger sister will be using AI, and I’m like, “Don’t do that. It’s bad for you.” Among everyone else in the world, I think that’s a rather staunch viewpoint. But there’s a small group of people in our generation who feel this way: if you are using AI, use as little as you can, or really try to work through it yourself.

Anika Kotapally: But I have some family friends who do use it, and when we talk about it, it’s very awkward in both directions. We’re very familiar with each other’s viewpoints, very unwilling to change them, and very aware of the flaws in both [viewpoints].

A Way in Through Fiction

 Shelley Lee: Given how you felt about AI, how did the two of you end up in a lab about it?

Chelsea Long: Through a short story. Our friend Yu Yan was in Professor Case’s History of AI class, and she shared “The IWM 1000” by Alicia Yánez Cossío with us, a story from 1975 that basically predicts the cell phone, and in a way AI. We all thought, “Whoa, this is super cool.” We went with her to Professor Case’s office hours to talk about it, and at the end Professor Case said, “Oh, we have this AI sci-fi reading group that you should join.”

Anika Kotapally: I was like, “That sounds fun. I’ll join.”

Shelley Lee: What made it something you wanted to keep coming back to?

Chelsea Long: It was my first experience being in an academic space that bridged the gap between undergrads and faculty. At one of the early meetings, we split into small groups, and Anika Kotapally, Yu Yan and I talked with a professor from the School of Public Health for about an hour. Ideas were just flowing.

Anika Kotapally: You do get that feeling of “You’re a faculty member. You understand things in a way that I don’t.” But in that conversation, I was like, “Oh, you’re very interested in what I have to say. You don’t think of me as an undergrad who doesn’t know anything.” It was a place where we could bounce ideas off each other despite our different levels of learning.

I also think our group has always been interested in humanizing our relationship with technology. Much of what we read imagined futures where people don’t lose their humanity in the face of technology, as we so often think they will, and instead gain things back because of it. It felt like a very hopeful space.

Informed, not Converted

Shelley Lee: What has changed for you over the last two years of learning and thinking through the lab?

Anika Kotapally: I wrote my final paper on how people conceive of AI today, a kind of mini present history of what people in a university think of it. I interviewed friends and read about how emerging technologies integrate themselves into our lives.

More than anything, that class opened my mind to what AI can be. Personally, I feel the same way I did: I don’t want to use AI, and I definitely don’t want to overuse it. I still think it’s bad for your thinking. But I’ve understood a lot more about what AI literally is and what it’s not, what it can do, and what it can’t or shouldn’t do. Throughout that class, I came to understand the limits of my own information. Now I feel way more informed and way more ready to talk about it. 

Chelsea Long: I have a more expansive conception of what artificial intelligence can look like. At the lab’s end-of-year lunches, where all the reading groups present, I heard what the other groups were doing: a neuroscience angle, math and language. It made me realize how many ways there are to approach AI through humanistic methods. People ask big questions there: What is a mind? What is intelligence? Is this intelligence really artificial at all? And a lot of what we read had AI or robot protagonists. You’re literally reading from the point of view of an artificial mind, immersed in characters with feelings and experiences that are real, or at least feel real. What AI does, all the mathy stuff under the surface, is really similar to how our minds make decisions. The more time I’ve spent with it, the fuzzier that line seems.

Anika, I was struck when you said you don’t feel under-informed. I think that’s one of the most important things about this lab. It gives people in the humanities comfort talking about AI, when we can often feel like interlopers, so we can engage on equal terms with more STEM-oriented people. And it brings the methods of the humanities to computer science and to STEM as a whole: reading, writing, talking, writing an essay. Approaching AI from that angle lets us understand it better.

Stochastic Parrots and Moving Goalposts

Anika Kotapally: The lab has been important for educating humanities people about what AI is. A common problem in the humanities, and I include myself until basically this spring, is that our understanding of AI starts and stops at the “stochastic parrot” idea, which was very popular on Twitter. [The phrase comes from a 2021 paper by Emily M. Bender, Timnit Gebru, Angelina McMillan-Major and Margaret Mitchell. Ed.] It basically means AI is just parroting back what you give it. It’s not thinking. It’s just selecting the next most likely word. It doesn’t know anything.

That was more correct than incorrect when AI was first getting big. Many humanities people glommed onto it, because it felt really good, when your entire job is about writing and thinking, to say that the writing-and-thinking machine can’t actually write and think, and that you’re the only one who can.

But AI progressed significantly, and very quickly, and its rate of improvement has only increased. Because a lot of humanities people turned their faces away from AI, it was easy to keep that phrase alive in their hearts, in my heart, even when it wasn’t true anymore. People are still stuck there, even now, when I think AI is leagues beyond what it was three years ago.

Every time AI improves, we move the goalposts on what being a person is, or what human thinking is. Scientifically, I think that’s bad. We can do that forever, and if we keep doing it, we’ll never understand the real differences and the actual similarities between human and AI thinking. Internally, in the humanities, it’s very useful to think of AI as worse than human, because we want to keep the human stuff to ourselves.

On the other hand, I think the role of the humanities is to bring the human to AI and force it into the conversation. A lot of STEM fields can be very rigid about being objective and uninterested in human perspectives, and the humanities are the exact opposite. What I’ve gotten out of this lab is learning to merge the two and think humanistically about AI. I think that also lets people like us help the people developing AI be better at their jobs, and better for the people who use what they make.

Knowledge as Output, Knowledge as Process

Shelley Lee: What you’re describing is something faculty and administrators are still working out. Many of us started where you did on AI [skeptical], though many of us haven’t spent much time with the tools themselves. We like hearing that the humanities will have a resurgence in the age of AI, but we’re still figuring out what that means for how we teach, and I suspect students have a lot to teach us. From where you sit, what should we be paying attention to?

Chelsea Long: AI has hit mathematics recently the way it hit the humanities in 2023 and 2024. OpenAI recently published something like 700 verified proofs. AI is doing the kind of mathematical thinking we thought only humans could do. It’s another moving-the-goalposts moment.

But Yu Yan sent me a fascinating article by a working mathematician. He argued that the output of mathematics as a discipline is not theorems and proofs. It is human understanding of those theorems and proofs. I love that lens. And it carries over to the humanities: the goal is not the output of papers, or of AI thinking, but human thinking and human knowledge. I also wonder whether it goes the other way. We use reading to study the sciences. Why can’t we use problem sets to study the humanities, in a way? There’s a big bridge between disciplines that needs to be built, and to keep being built.

Anika Kotapally: AI hit chemistry very early, because it has a lot of useful applications there. I’m going to posit this now: there are two different types of knowledge output. One is where what we want is the end goal, the answer.

Chelsea Long: Solve the whole problem.

Anika Kotapally: Yes. And AI is very good at that. AlphaFold is revolutionary and will change medicine and lives forever, because as long as the answer’s right, we don’t care how AI gets it. Biology and chemistry have used mathematical models that way for a long time. As long as someone can make a model that works, we don’t care what’s happening in the middle.

The other way, which you just described very well, is knowledge as… 

Chelsea Long: As process?

Anika Kotapally: Yes, as process, not as output. I think that’s true for math and for the humanities. For a lot of humanities majors, college is not about which papers you write. It’s about the processes you learn. You come to college to learn how to think, how to read, and how to write. They make you read theory so you learn how to read theory, not necessarily so you know the specific theory you read.

So going forward, I think that distinction may be more useful to higher education than the traditional STEM-humanities divide. If you’re in a field where knowledge is output, AI is very useful to you. But if you’re in a field where knowledge is process, AI is not useful to you at all, because it cannot explain a process to you in a way that is useful. Sure, it did those 700 proofs…

Chelsea Long: [It] doesn’t mean you understand them.

Anika Kotapally: Right. It doesn’t mean anything, because we don’t understand how it did it, and we want to understand how it did it so we can do other things with that information. Maybe that is what higher education can take away from this lab. The distinction itself is humanistic. It’s concerned with what we know, how we know what we know, and what we want to know.

Shelley Lee: That distinction could change the goals of teaching. In history, we assign papers based on readings, but now you can feed those readings into a machine and have it spit out an analysis in a few seconds. So the focus has to shift to learning and understanding: mulling over ideas, working them over, talking them through. I put it this way to a group of students I was talking to: “You can now get your degree and have a machine do everything for you.” They said, “Then what would be the point of being in college?” Exactly!

[Anika and Chelsea nod]