But it still doesn’t feel like closure.
The origin story of L.A.’s Palisades Fire, according to a criminal complaint announced yesterday, reads like a scene from an art-house film. Shortly before midnight on New Year’s Eve, a son of missionaries visits a scenic overlook near the Los Angeles coast. The clearing is known for the Buddha statues hikers leave behind in the hollowed-out stump of a power pole. The man listens to a French rap song about the malaise of modern life. Then, according to investigators, he starts a fire with an open flame, a combustible material, and malicious intent.
He dials 911 to report the fire, but his first few calls do not connect (presumably because this is coastal Los Angeles, and our cell service is terrible). He then begins screen-recording on his cellphone while he continues to dial 911. He asks ChatGPT if he might be criminally liable for starting a fire with cigarettes, possibly to cover up what he’s done. Then, the man films the flames on his iPhone as firefighters arrive.
By January 2, they determine that the fire is out. But it has in fact gone underground, smoldering in the root system of the hillside’s brush. Days later, strong winds travel from the desert to that same hillside and revive the blaze, which becomes the Palisades Fire. It levels more than 6,800 structures and kills 12 people. (Those structures included my childhood home, and those deaths included Arthur, a man who’d lived next door to that old house and whom I’d known and loved since I was born.)
The suspect is a 29-year-old Florida man named Jonathan Rinderknecht, and the case against him is one that could be made only in an era of AI. To help establish intent, the Bureau of Alcohol, Tobacco, Firearms and Explosives turned to Rinderknecht’s conversations with ChatGPT—not just his cigarette question, but also an exchange from months earlier in which he asked ChatGPT to generate an image of a “burning forest” next to a crowd of people “running away from the fire.” OpenAI declined to specify whether the company had handed Rinderknecht’s chat logs over to the investigators; a spokesperson for OpenAI told The Atlantic only that “following the Palisades fire tragedy, we responded to standard law enforcement requests related to this individual.
