Culpability
Culpability
A modern alternative to SparkNotes and CliffsNotes, SuperSummary offers high-quality Study Guides with detailed chapter summaries and analysis of major themes, characters, and more.
Summaries & Analyses
Author Insights
NEWQuizzes
NEWReading Tools
Bruce Holsinger on Culpability
In him own words: quotes and insights from the author.Why he wrote it
A helicopter, a helipad, and years of waitingThe setting came before the story. During the first fall of the pandemic, Holsinger kayaked out of a cove on the Northern Neck of Virginia and came upon something that didn’t fit: a sprawling private estate that had swallowed what had been a horse farm, complete with a fake lighthouse and a white sandy beach. He watched a helicopter land on a helipad. Nothing followed. The image stayed dormant for two years before any plot attached itself to it.
I went and kayaked out that first day and passed at the end of the cove at the point, this, what looked like a really out of place, for lack of a better word, estate or compound, exactly as described in the novel. It was, just the trees had been raised, there were tennis courts, there was this massive renovated house, there was this fake looking lighthouse thing and this white sandy beach.
The location and the AI premise did not arrive together. The Northern Neck sat in his imagination as a setting in search of a plot, and the moral architecture of the novel only snapped into place once the technology caught up with what his story needed: a car that could share responsibility for what happened.
Culpability had two points of origin: a family outing to the Northern Neck of Virginia, where I was initially inspired to set a novel by the Chesapeake Bay; and the sudden mania for Artificial Intelligence beginning in late 2022, when ChatGPT came on the scene. I was resolved to set a novel in that location, and it was only gradually that the AI and moral responsibility themes got layered into the book as part of my writing process.
The structural question that made AI necessary was: who was actually driving? Once he decided the car drove itself, the novel’s central problem shifted from a teenage boy’s recklessness to something far harder to prosecute or forgive. He finished a full draft before realizing the technology had changed the book’s subject entirely.
The initial spark that drove me into thinking about AI was that I wanted their car to be a self-driving car so there would be a nonhuman aspect of responsibility that could be explored. I finished a draft in the fall of 2022, and at that point ChatGPT had just blown up. That’s when I realized the novel was about artificial intelligence, agency, responsibility, and the way those categories interlace with this family going through a calamity.
His interest in AI was not neutral curiosity. As an English professor watching large language models reshape student work from inside a university classroom, he came to the subject with something closer to alarm than fascination.
I’ve been fascinated and repulsed by the propensity of AI to threaten human autonomy. Corporations are building these machines while humans are giving more and more of their everyday lives and decision-making over to their algorithms.
How he wrote it
One voice, many documents, one late additionEvery earlier Holsinger novel had spread its narration across multiple third-person perspectives. For this book, a trusted beta reader pushed him toward first person after reading an early draft. He made the switch, and it changed how the novel generates suspense: the reader is locked inside a consciousness that consistently misreads the people closest to it.
This is the first time I’ve written one all the way through in first person, really, from one person’s point of view.
The narrator, Noah, is not dishonest. The gap between what he notices and what is actually happening in his household is not a trick but an architecture: it is where the novel’s tension lives. Holsinger drew a careful distinction between two different kinds of unreliable perception when describing Noah on the WAMC Book Show.
I wouldn’t call him an unreliable narrator, but he is a somewhat oblivious narrator. He’s in some ways blind to what’s going on around him with his kids, with his wife. [paraphrased from audio]
First-person narration creates a confinement problem: how do you let the reader see beyond what one consciousness registers without breaking the frame? Holsinger’s solution was to build a set of documents into the novel that exist outside Noah’s account: his daughter Alice’s chatbot exchanges, fragments of his wife Lorelei’s book-in-progress on AI ethics, and a mock New Yorker profile of a tech billionaire. These carry plot-critical information that Noah never directly observes. The revised opening was another structural correction. He had originally buried the accident in backstory, but recognized that structure wouldn’t hold.
I really want to start with the accident. So instead of getting that as backstory, you get it right up front. You get it in the first chapter. [paraphrased from audio]
The character who most required research did not exist in the first draft. Lorelei, Noah’s wife, is a professional AI ethicist whose book-in-progress runs through the novel as a series of interpolated fragments. She materialized only once Holsinger understood that the novel’s argument needed a character with formal expertise in the very systems the plot turns on. To write her credibly, he read across a field he had no prior specialist knowledge of.
I read widely in research on ethical AI, algorithmic inequity, and related topics, including the work of Fei-Fei Li, Timnit Gebru, Thilo Hagendorff, and many others. I’m not an expert in the topic, but I learned enough to be able to sketch Lorelei’s life, profession, and work in a way I hope is convincing to readers.
The research also surfaced something he had not expected. Talking to AI researchers and a UVA law school colleague working on algorithmic bias, he kept encountering a practice he found both compelling and chilling: technologists casually assigning a numerical probability to the chance that AI would end humanity.
That discovery made it into the novel through Lorelei’s voice, but it also shifted his own sense of what the book was documenting. The term, used routinely inside AI research communities, arrived in his understanding of the field as an index of how seriously the people building these systems take the possibility of catastrophe.
I was intrigued and horrified to learn that many AI researchers — the technologists, engineers, programmers and ethicists — will often tell you their ‘p-doom’ number: a number on a scale of 1 to 10 or 1 to 100 that reflects the probability that AI will extinguish humanity (though given the environmental degradation involved in feeding data centers with water and power, the threats of AI and catastrophic climate change go hand in hand).
What he wants us to sit with
Culpability as a shared condition, not a verdictThe title does precise philosophical work, and Holsinger chose it over his original working title, “Autonomy,” because it places the weight differently. Culpability names a diffuse mode of responsibility rather than a legal finding, and it can attach to a family, a corporation, and an algorithm simultaneously. He reached for a distinction the law makes but everyday speech tends to blur.
You can be culpable without being guilty. You can be guilty without being culpable. So, I wanted that slightly ambivalent sense of our responsibility for what we do and don’t do. [paraphrased from audio]
The novel does not argue a case for or against AI, and Holsinger resists readings that turn it into one. He has a twenty-year academic career in medieval literature behind him: he knows the difference between a text that persuades and a text that opens a question. The distinction matters to how he thinks about what fiction is for.
A novel is not an op-ed, and I try to think of my novels as thought experiments. If there’s a philosophical takeaway as it relates to AI, it would be that we need to take a beat, pause, and think about what we’re doing. We should embrace these increasingly elusive moments that are free from technology.
The most visceral version of his argument against AI’s logic came not from his research but from a question asked at one of his book events. A lawyer in the audience wanted to know whether AI tools might help him write prose faster. The question landed as a category error, and his reaction to it articulates the novel’s deepest concern more directly than any passage in the book.
My head almost exploded. The thought of using AI to write prose … one of the pleasures of writing is in the lack of efficiency! I have no interest in efficiency when it comes to writing. I think that’s one of the scariest dangers of AI — that it comes with the notion that everything should be efficient, when a lot of parts of life maybe shouldn’t be efficient.