The Algorithm is Gonna Get You

Privacy, Surveillance, and the Palisades Fire Arson Trial

Federal prosecutors allege that Pacific Palisades fire suspect Jonathan Rinderknecht used ChatGPT to generate images of burning landscapes; image shared on X by First Assistant US Attorney Bill Essayli.

There is a particular kind of exhaustion that sets in when the news cycle has ceased feeling like a sequence of events and starts feeling like a condition. Somewhere between the gas prices and the geopolitical whiplash, between the UFC circus on the White House lawn and the threat of New World screwworm, it can become difficult to sustain attention for any individual story long enough to think about it seriously.

Which is perhaps why an unfolding story about privacy, surveillance, big tech, and civil liberties has not gotten quite the attention it warrants. In October 2025, federal authorities announced the arrest of Jonathan Rinderknecht, a 29-year-old Florida man (really!) charged with igniting the Lachman Fire the previous New Year’s Eve — the blaze that smoldered underground for six days before resurging as the Palisades Fire, which killed twelve people, destroyed nearly 7,000 structures, and burned over 23,000 acres in what became the most destructive fire in Los Angeles history.¹ His trial began earlier this month. The case is, on its surface, a story about arson. But the means by which investigators built their case against Rinderknecht is yet another story about how comprehensively, and how unremarkably, we are all being recorded.

Investigators assembled a detailed account of Rinderknecht’s movements and state of mind without his cooperation, and largely without his awareness. GPS data, Uber logs, cell tower records, surveillance camera footage, and the contents of his ChatGPT conversations allowed prosecutors to reconstruct not just where he was on the night of December 31, 2024, but what he had been thinking about for months. Prosecutors presented evidence that he had used ChatGPT to complain about wealth disparity and corporate greed, had prompted it to generate an image of a dystopian burning city, and had vented into it in real time in the hours before the fire, simultaneously texting a romantic interest and typing his frustrations into a chatbot that was keeping meticulous records. The servers kept the score.

What makes this case notable is not simply that digital evidence was used in a criminal investigation — that has been routine for decades. It is the particular character of what was captured. Phone records tell you who someone called. Search histories tell you what someone looked up. But a ChatGPT conversation log is something substantially different: it is a record of thought in progress, of the private reasoning and feeling that people have historically conducted in their heads, in diaries, or in conversations they assumed were ephemeral. Rinderknecht did not know he was keeping a dossier on himself.

But the ChatGPT logs were only part of the picture. What prosecutors assembled was not a single surveillance record but a mosaic: GPS coordinates placing Rinderknecht at the trailhead, Uber’s timestamped ride data, cell tower pings, surveillance camera footage capturing his vehicle, DNA from a lighter found in his car. No single data point was necessarily conclusive. Taken together, they produced something close to a continuous record of one person’s movements, relationships, emotional state, and private thoughts across months. This is the aggregation problem that privacy scholars have been warning about for years, and the Rinderknecht case is one of the clearest illustrations of it playing out in a courtroom: the whole is not just greater than the sum of its parts, it is of a different order entirely.

---

It is also worth noting what Rinderknecht did not do. He did not consent to being surveilled. He did not post a public manifesto. He did not hand over his phone. He went about the ordinary business of contemporary life — driving for Uber, using a navigation app, talking to a chatbot — and in doing so generated a comprehensive record of himself that was waiting, as it turned out, to be subpoenaed. This is the condition all of us are in, more or less, all of the time.

For librarians and information professionals, this should not be an entirely unfamiliar dynamic. The question of what records institutions keep, who can access them, and under what circumstances has been central to library ethics for over half a century — and the profession’s engagement with it has not always been comfortable. In 1971, the American Library Association adopted its Policy on Confidentiality of Library Records, strengthening protections for patron data at a moment when the FBI's interest in library borrowing habits was an open secret in the profession if not yet in the public record. That instinct proved well-founded: when the Church Committee investigations commenced in 1975, they confirmed that the FBI had been systematically monitoring the reading habits of Americans it deemed politically suspect, requesting borrower records from libraries as part of broader surveillance programs. Many states subsequently passed laws protecting patron records. The underlying principle was straightforward — what a person reads, searches for, or thinks about is their own business, and institutions that facilitate that inquiry have an obligation not to become instruments of the state.

The passage of the USA PATRIOT Act in 2001 reopened that wound with considerable force. Section 215 gave the FBI authority to demand library records — borrowing histories, internet search logs, any “tangible things” deemed relevant to a terrorism investigation — without a warrant and with a gag order preventing librarians from disclosing the request had ever been made. The American Library Association responded forcefully, passing a resolution in 2003 declaring Section 215 a present danger to the constitutional rights of library users and encouraging libraries across the country to post notices warning patrons that their records were no longer safe. Attorney General John Ashcroft’s response was to accuse the library community of spreading “breathless reports and baseless hysteria.” He also repeatedly insisted the government had never served a national security letter on a library. And, well, that turned out to be untrue. In 2005, the Library Connection, a consortium serving thirty libraries in Connecticut, received exactly such a letter — and its directors, gagged by law, had to watch Congress debate the Patriot Act’s reauthorization without being able to say a word.

The profession was right to be alarmed. And it was right for a reason that extends well beyond libraries: the principle that the state should not have easy access to records of what people read, search for, or think about is not a librarian’s eccentric preference. It is a cornerstone of intellectual freedom in a democratic society.

Which brings us back to Jonathan Rinderknecht, and to the rest of us. The library community spent decades fighting to limit what records institutions could be compelled to hand over to the state. It won some of those fights. Many states have strong patron confidentiality laws. The ALA’s position on privacy remains unambiguous. But those protections were built for a world in which the most sensitive record of your intellectual and emotional life was held by a library or a bookstore — institutions with professional ethics, legal obligations, and at least some incentive to push back. That world has not disappeared, but it has been substantially overtaken by one in which the most comprehensive record of your inner life is held by a technology company whose terms of service you likely agreed to without reading, whose data retention policies you probably never looked up, and which has no professional canon obligating it to treat your privacy as a value rather than a liability.

When Rinderknecht typed into ChatGPT, he was not thinking about subpoenas. He was thinking about whatever he was thinking about — loneliness, grievance, the state of the world. That is what these platforms are designed to elicit. They are built to be confided in, to feel like a private space, to absorb the kind of unguarded thinking that people have historically done in their heads or on paper they could burn. The intimacy is the product. And the record of that intimacy is, it turns out, available to the government under circumstances that are considerably less restrictive than those that apply to your library account.

None of this requires us to feel sympathy for Rinderknecht, who has pleaded not guilty but faces serious charges in connection with a genuine catastrophe. The point is not that investigators shouldn’t have access to evidence. The point is that most of us have not reckoned with what we are generating simply by living our lives — driving, navigating, ordering things, talking to chatbots — and what that record looks like in aggregate to someone with a subpoena and a reason to look.

The good news is that challenging panoptic surveillance is not uncharted territory. I touched on the role of the American Library Association and the work of librarians. Information professionals and like-minded scholars have produced frameworks, resources, and arguments that extend well beyond the library world. My Heterodoxy in the Stacks colleague Sarah Hartman-Caverly, a librarian and information ethicist at Penn State Berks, has spent years writing and teaching at exactly this intersection: privacy literacy, surveillance, and what it means to exercise genuine intellectual freedom in an age of pervasive data collection. Her contributions to this Substack are a good place to start if you want to go further than this essay takes you; and I highly recommend exploring the Digital Shred Privacy Toolkit, a comprehensive resource co-curated by Sarah and her Penn State Berks colleague Alexandria Chisholm.

---

We have, in the span of roughly fifteen years, incidentally constructed a surveillance infrastructure that would have been the envy of every attorney general who ever sent an agent to a library looking for a reading list. We did it because the products were useful and free and frictionless, and because the nothing-to-hide argument is surprisingly durable even among people who should know better. The Rinderknecht case is not a cautionary tale about one troubled man’s digital footprint. It is a window into what already exists for all of us — waiting, indexed, and available to anyone with a lawful reason to look.

The legal scholar Daniel Solove has argued that the real danger of ubiquitous surveillance is not Orwellian but Kafkaesque — not the fear of being watched, exactly, but the powerlessness of knowing that somewhere, a record exists, that it can be retrieved and interpreted and acted upon, and that you will have no meaningful role in that process. Rinderknecht did not know he was composing a portrait of himself, yet the portrait exists anyway, and it has been presented to a jury. Most of us are composing similar portraits, continuously, without knowing it. The question is not whether we have anything to hide. The question is who gets to decide what the portrait means.

---

To promote viewpoint diversity, Heterodoxy in the Stacks invites constructive dissent and disagreement in the form of guest posts. While articles published on Heterodoxy in the Stacks are not peer- or editorially-reviewed, all posts must model the HxA Way. Content is attributed to the individual contributor(s).

To submit an article for Heterodoxy in the Stacks, send an email with the article title, author name, and article document to hxlibsstack@gmail.com. Unless otherwise requested, the commenting feature will be on. Thank you for joining the conversation!

References and Further Reading

American Library Association, “Policy on Confidentiality of Library Records,” adopted January 20, 1971.

Boston Public Library, “Historic Congressional Committee Hearings and Reports: Church Committee (Intelligence Activities, 1975),” research guide.

Alexandria Chisholm and Sarah Hartman-Caverly, “Digital Shred Privacy Toolkit.”

Uwa Ede-Ofiso, “Accused LA Wildfire Arsonist Wanted ‘Revenge on Society’, Prosecutors Say as Trial Opens,” The Guardian, June 10, 2026.

Jack Hannah, “A Vengeful Arsonist or a Convenient Scapegoat? What We’ve Learned from Testimony in the Palisades Fire Trial,” CNN.com, June 21, 2026.

Sarah Hartman-Caverly, “The Secret Ingredient In Civic Dialogue Education,” Heterodoxy in the Stacks, February 10, 2026.

Josh Haskell, “29-Year-Old Man Accused of Starting Blaze that Turned into Destructive Palisades Fire,” ABC7.com, October 8, 2025.

June Kronholz, “Patriot Act Riles An Unlikely Group: Nation’s Librarians,” Wall Street Journal, October 28, 2003.

Tim Marcin, “Dystopian ChatGPT Image Helped Catch Alleged Arsonist in Deadly Pacific Palisades Fire,” Mashable, October 9, 2025.

Daniel Solove, “‘I’ve Got Nothing to Hide’ and Other Misunderstandings of Privacy,” San Diego Law Review, 2007.

Daniel Solove, “Why Privacy Matters Even if You Have ‘Nothing to Hide’,” The Chronicle of Higher Education, May 15, 2011.

Maia Spoto and Vik Jolly, “Accused Palisades Arsonist Resented the Wealthy, Prosecutors Say,” New York Times, June 10, 2026.

US Attorney’s Office Central District of California, “Florida Man Arrested on Federal Criminal Complaint Alleging He Maliciously Started What Became the Palisades Fire,” October 8, 2025.

US Attorney’s Office Central District of California, “United States v. Jonathan Rinderknecht,” updated June 18, 2026.

Deborah Walters, “15 Years of Fear: The ALA and the Patriot Act,” ALA Archives blog, October 7, 2024.

Tian Wang et al, “Patron Privacy Protections in Public Libraries,” The Library Quarterly 93 no. 3, July 2023.

Steve Witt, “The Evolution of Privacy within the American Library Association, 1906–2002,” Library Trends 65, no. 4, 2017.

1. For scale, consider that New York’s Central Park is 843 acres — so, roughly 27 Central Parks burned. Or, for a California-centric example, consider that San Francisco is about 30,000 acres.