Vauhini Vara: On AI’s Place in Literature, Writing into the Zeitgeist, and Her Newest Book ‘Searches: Selfhood in the Digital Age’

I first became familiar with Vauhini Vara’s work after reading her viral essay “Ghosts” in The Believer. “Ghosts” is a short story that integrated responses from an early version of ChatGPT into the story. An introduction precedes it which says, “In the nine stories below, I authored the sentences in bold and GPT-3 filled in the rest. My and my editor’s sole alterations to the AI-generated text were adding paragraph breaks in some instances and shortening the length of a few of the stories; because it has not been edited beyond this, inconsistencies and untruths appear.”

The story was published in the summer of 2021, more than a year before ChatGPT was publicly launched. In the years preceding ChatGPT’s launch there were multiple soft launches that had many people hypothesizing about how large language models might shape the way we interact with the world. A question that regularly emerges is: How will the use of LLMs impact the way we consume and make art and literature?

Now, three years after the publication of “Ghosts,” Vauhini Vara has published Searches: Selfhood in the Digital Age (Pantheon, 2025), which builds on the work of her viral autobiographical essay and her years of working as a journalist covering tech. Searches gives a clear-sighted perspective on a subject that continues to be controversial and elusive. Interweaving in-depth accounts of how companies like OpenAI and Meta have shaped our culture with more personal sections that showcase how Vara has engaged with technologies as a person and a writer, Vara has created a multifaceted portrait of artificial intelligence in our modern world. She might be the only person who can write about tech bros and their toys in a way that doesn’t make me fall asleep, which might be a result of masterful storytelling skills that somehow manage to make informational narratives feel a little bit like gossip. Tying it all together are chapters in which Vara has conversations with a chatbot about the book itself.

The result is a book which is difficult to categorize, and which resists easy answers to complex questions. In an age where technology is working to mitigate the need for deep analysis or contemplation, Searches gives its reader the gift of ambiguity and invites them to firmly maintain the human ability to engage critically and curiously if one so wishes to do so.

I spoke with Vauhini via Zoom in late March about Searches: Selfhood in the Digital Age, her relationship with technology, and AI’s place in literature. 

Shelby Hinte: The conceit for Searches feels really original. I love the way the chapters/essays are kind of fed to AI two at a time. And then we see you, the writer, in conversation with AI. Were those all real conversations with AI?

Vauhini Vara: Oh, yeah. But the thing that I want to make clear is that I wrote the whole book outside of that AI conversation and then later fed it chapters two at a time. So I wasn’t really asking for feedback that I was going to implement in the book. I think of those interactions with ChatGPT as being, in some way, like a power game. It was manipulating me, but I was also manipulating it.

SH: How did you come up with that idea to incorporate the AI conversations into your book in that way?

VV: So, to give credit where it’s due, I had two editors for this book because the book was acquired by Lisa Lucas at Pantheon, and then she left Pantheon and Denise Oswald became my editor right in the middle of the editing process. So I actually got rounds of edits from Lisa, and then the book moved to Denise, and I got a round of edits from Denise. But when I first turned in the book it was actually just the even-numbered chapters as they exist in the book now, which were essentially experimental essays that included language from my interactions with tech companies’ products. And Lisa read that and two of her big pieces of feedback were, number one, “I think you know what the subtext is in these essays, but I don’t know that readers necessarily will. Would you think about maybe adding an introduction and conclusion so you can set it up for readers?” And I thought about that and was like, you know, she makes a really great point, but what I think I need is actually an additional throughline in the book. And that’s where all the odd-numbered chapters, which create this kind of narrative arc from beginning to end, came from. 

Another thing she said was that ChatGPT is such a big part of the discourse about tech now, but it wasn’t that present in the version of the book she had. She said something like, “I wonder how you might engage with ChatGPT; I wonder what would happen if you even fed parts of the book to ChatGPT.” My immediate reaction was like, “Oh, what a dumb conceit.” I didn’t take it seriously at first. And then I realized that engaging with ChatGPT could actually allow me to do with ChatGPT something similar to what I’m doing with my engagement with other tech products that show up in the book, like my Amazon reviews and my Google searches. So I tried it and that’s when I realized that, in addition to revealing certain things about me as a narrator, my use of ChatGPT revealed some things about the technology and how it works, too.

SH: It’s so interesting to hear that because at the beginning of the book, we start with an essay that’s a little bit more narrative and then we get to the reviews and the searches, which I felt were so experimental, and I had this moment where I laughed out loud because I remembered, you know, halfway through reading one of the lists, the narrator just fed this material to AI. What is AI going to make of this? Will it be totally baffled? And then I was shocked to see that it actually responded in a semi-articulate way to the more experimental pieces.

VV: Right? Like it kind of “understood” what was happening there.

SH: Yeah. Did that surprise you as well? What was your response?

VV: That’s a good question, because sometimes when I gave it a more experimental chapter, it sort of understood what the experimentation was about, and then sometimes it didn’t. 

The question of surprise is a hard one to answer, I think because with these large language models they’re new enough that I never quite know what to expect. And that’s part of what’s interesting and exciting in some ways. It’s also what’s dangerous and threatening in others. 

SH: Later in the book, you really touch on some of the biases of the tech—it excluding women or people of color in certain pieces of information that it would feed you, for one. I had a pretty visceral response to the moment in the book where the AI created revised passages of the original Sam Altman pages you fed it. Not only were they entirely different in tone from the original, but they also included a huge increase in details that felt unrelated to the original text. What was your impression of that when you were engaging with the AI?

VV: Part of me hesitates to say, because I find it so delightful that any reader can read that section differently. I never asked AI to write a revision or provide text for me about Sam Altman. The chatbot just went ahead and did that. So lots of things are happening in that. First of all, it’s meant to be a nonfictional text, and the chatbot is inserting fictional details. Secondly, it is writing a text about Sam Altman, the CEO of the company behind the product ChatGPT, that is much more credulously positive than anything I would write. And it’s doing this without announcing that it’s doing it. So it requires a fair amount of critical thought on my part as the person who’s engaging with ChatGPT to see what’s happening there. And you know, I’m somebody who is a long-time tech reporter, who is a writer, who knows how to engage with literature, but that’s not necessarily a skill that everybody using ChatGPT will bring to it. I’m hoping that showing these things through my engagement with ChatGPT will draw attention to some of the subtle things these technologies can do that we may not even be aware of.

SH: One of the most hilarious things to me in the book is when you ask ChatGPT to tell you who you are and then it provides misinformation, but it does so with such an authoritative tone. In a later conversation the AI refers to this tone as implied objectivity, which isn’t that different from what we’ve been trained in school to believe the news is. Do you think there’s any writing that can ever be objective? 

VV: I would venture to say that every communication act is subjective. There are a lot of forms of communication that are assumed to be objective, but which are clearly subjective, including, you know, news articles in the New York Times. It’s not only an author’s subjectivity that influences the article, but it’s their multiple editors’ subjectivity, the subjective institutional perspective of the publication. So by that logic, there’s no objectivity behind the language that ChatGPT produces or that any AI model produces, which then leads to a couple of questions, one of which is, what are the subjective perspectives that are being used to train these AI models and inform the language that they’re producing? I think an equally relevant question is, what is the institutional perspective of the companies behind these products? 

SH: You say something really lovely in the book about hoping AI can give you words where you maybe previously didn’t have words to say the thing you wanted. I thought about that specifically with writer’s block, but even more specifically with your story “Ghosts,” which you’ve written about the difficulties in finding a way to write that story because of the subject matter. In my email to you, I told you I’ve been teaching that story in my classes pretty much ever since I first read it. And one of the main prompts that I ask my students to do with it is to write the thing that is too hard to write. This feels like a major theme of your book—trying to write the thing that feels impossible to write. Why does that challenge feel important to you?

VV: I started climbing recently, within the past few months. I’m not great at it, but I’m really enjoying it. When I started climbing, obviously there were certain routes in the gym that seemed impossible to me and certain routes that were easy and certain ones that felt like a challenge that I was comfortable with. And from the beginning I always gravitated toward the ones that were comfortable but felt challenging. And then once those ones felt easy, I moved on to a level that again felt comfortable but challenging. And in that way I eventually found myself being able to comfortably climb those routes that, when I started climbing, looked really difficult, or almost impossible to me. 

And the reason I bring that up is because I think there are lots of things for all of us where we naturally gravitate toward that which is more difficult rather than less difficult, because it’s more fun and it’s more interesting.

This metaphor can extend and extend, because the other thing that’s interesting about the routes  that are comfortable but challenging is that you reach a point on the route where it is really, really hard for you, and you might have to do the same one three times or a dozen times in order to get past it. And then when you do, that’s actually more interesting and satisfying than getting to the top of it. They’re called problems, and figuring out a problem is at the heart of what makes climbing fun and interesting. And I think that’s true for writing, too. I think what makes writing fun and interesting is that once I’ve figured out how to do a particular thing, I want to try to do something harder. 

I think writing, and communicating in general, is always difficult—from the smallest kind of communication to the most complex. I think any of us who are writers are drawn to that challenge. There’s something that feels emotionally satisfying about being able to convey the subjective reality of human experience.

SH: Yeah, it makes me think of your essay, “I’m Hungry to Talk,” which is about your experience speaking Spanish. I’ve been practicing Spanish on and off for years and I had this realization a while back that I don’t have a sense of humor when I speak in Spanish. I am not able to make jokes. My favorite thing is to make my friends laugh, and to think that a fundamental part of me is lost, or would be lost in certain circumstances, is kind of painful. You really capture that experience in “I’m Hungry to Talk,” and I loved the moment when you quote an acquaintance of yours who says, “When it comes down to it, all communication is a risk of involving misunderstanding.” How do you think that awareness of the fallibility of language shapes the way you write or what you choose to write about?

VV: The desire for human-to-human connection and the ways in which our ability to fulfill that desire falls short is a subject that I find endlessly interesting—whether it’s in sibling dynamics or parent-child dynamics or romantic partner dynamics or friend dynamics. That’s something that I find fascinating and so I think it feels so bound up with my interest in writing. 

SH: When do you invite other people in to read your work and share feedback? 

VV: The funny thing is, normally I will get feedback starting pretty early on. For my novel and my story collection, I had friends or my husband—who is also a writer—read those multiple times and give me feedback. And then I would revise and then ask the same people, mostly my husband, to just read it over and over. I find that really important. The thing about Searches is that I sold it on proposal to Lisa Lucas and I was kind of racing to write it. I didn’t have time to show it to anybody. So the first person who saw it was Lisa. The second person who saw it was Denise Oswald.

I never really got early feedback on any of this book. And so, this is the book that sort of burst forth from my head onto the page with not a lot of integration of feedback. For what it’s worth, it may also be the case that since it’s my third book I have honed that internal sense of what others might say and also honed my ability to then respond to that. So, this one was different, but it’s not the approach that I prefer. I always prefer to get feedback rather than not.

SH: What did it feel like to write something like this when in the past you had the habit of having so many drafts that you shared and received feedback on?

VV: I was really nervous. I feel better now because some early readers have read it and responded positively. But before I felt really nervous. The question was, like, I think I’ve written a great book, are other people going to agree? I had no way of knowing, and, in my view, the book doesn’t really exist until people have read it. That’s what makes the book come into existence. It is not just the fact that I wrote it. The magic happens when a reader reads a book and a spark takes place between author and reader. 

SH: Do you think any of the nervousness was coming from the risk of writing into something that is such a hotly debated topic in our current zeitgeist?

VV: I was trying to write a book that talked about technology and our relationship with it in a very non-binary way, in a way that wasn’t just saying technology and these products are horrible and these companies are out to get us. It also wasn’t saying these companies are amazing and these products are amazing and the companies are just trying to help and support us. It is a different argument that requires readers to hold both thoughts in their brains at the same time. I wasn’t sure how successful that would be. I wasn’t sure if I was conveying that on the page in a way that was going to be legible to readers.

SH: I think it can be nerve-wracking to put any opinion out there where everyone can so publicly comment on it. And yet that’s part of the work of being a writer to some extent—giving opinions and sharing perspective. When you’re writing into something that’s controversial, how do you find the confidence to write the work that you feel most inspired to write?

VV: That’s actually another thing I was nervous about. Because my background in writing about technology is as a journalist, it’s important to me to be fair and fact-based in my work. So constantly with this book, I was asking myself whether I was being fair and fact-based. Oftentimes, if there was an opinion that was really compelling to me but that didn’t seem fact-based, I would cite a theorist or a scholar who holds that opinion rather than citing it as fact on the page. I think opinion writing is valuable. I think the reason I feel a little uncomfortable choosing an opinion and putting it on the page is because of my journalism background. I think the facts as they exist can support all kinds of opinions. So I tend to feel like I value facts over whatever opinions are formed from analysis. I don’t want to overstate that because clearly there is a point of view in this book that I hold. It’s not the kind of text that I would’ve written in a newspaper article for the Wall Street Journal when I was covering tech companies for them. I’m still grappling with what that means.

SH: I don’t want to assume any obsessions of yours, but you’ve made a career of writing about technology both in your fiction and your nonfiction. What role do you think fixations or obsessions serve in the life of a writer?

VV: I think they’re great. In writing, the rewards are so often intrinsic rather than extrinsic. The internal rewards end up feeling more meaningful than the external ones—which is to say that obsessions give us a reason to write for its own sake. I think all of us who write are driven by obsession in some way. We write because we want to understand something. The thing we want to understand is typically something that we’re preoccupied by or obsessed with. I think the question when it comes to publishing is, can I be doing something new if my obsession is consistent? For me, the answer to that question is yes.

*

Vauhini Vara is the author of Searches, named an anticipated book by The New York Times, Esquire, Foreign Policy, and others, which Publisher’s Weekly called a “remarkable meditation.” Her previous books are This is Salvaged, which was longlisted for the Story Prize and the Mark Twain American Voice in Literature Award, and The Immortal King Rao, which was a finalist for the Pulitzer Prize and was shortlisted for the National Book Critics Circle’s John Leonard Prize and the Dayton Literary Peace Prize. She is also a journalist, currently working as a contributing writer for Businessweek, and an editor, most recently at The New York Times Magazine. She teaches at the Lighthouse Writers Workshop’s Book Project.