Prosecutors in the Palisades fire arson trial reached for an unusual piece of evidence: the defendant’s ChatGPT logs. According to The Verge AI, the case against Jonathan Rinderknecht ended in a hung jury and a mistrial after jurors split 10-2 in favor of the defense. He had been charged with setting a fire on New Year’s Day 2025 that grew into one of the deadliest wildfires in LA history.
What stands out here isn’t the arson case. It’s that someone’s casual conversations with an AI chatbot showed up in a courtroom as supposed proof of intent. And the jury didn’t buy it.
What prosecutors brought to court
The state built its case on the kind of evidence you’d expect. The Verge AI reports that prosecutors used location data from Rinderknecht’s iPhone, security camera footage, and witness testimony. Then they added his ChatGPT history to the pile.
According to the report, prosecutors said Rinderknecht:
- Had ChatGPT generate images of fire
- Asked the chatbot, “Why am I so angry all the time?”
- Ranted to it about how the wealthy were destroying the world
- Was caught on a screen recording asking whether someone could be blamed for a fire started by their cigarette
The argument was that these exchanges revealed a mindset. A motive. A character flaw, even.
Why the jury pushed back
The jurors weren’t convinced, and one of them said so bluntly. As detailed in The Verge AI, a juror told CBS LA that she didn’t think the ChatGPT logs proved anything, adding, “I talk to ChatGPT all the time.” She said the suggestion that using a chatbot pointed to some defect in his character actually made her “angry.”
That reaction matters more than it might first appear. The prosecution was betting that a jury would read AI chat logs as a window into someone’s soul. Instead, the jurors read them as something ordinary. People ask chatbots strange, dark, and venting questions every day. Treating that as incriminating felt like a stretch to the people in the box.
Why this matters
This is significant because it’s an early test of how AI conversations play in front of a jury, and the answer was skepticism. Your search history has been fair game in court for years. ChatGPT logs are a newer frontier, and they’re different in a key way: people talk to chatbots conversationally, the way they’d talk to a friend or a therapist or themselves. That makes the logs feel intimate. It also makes them easy to misread.
A few things worth sitting with:
- Chat logs are discoverable. If prosecutors can request them, your conversations with an AI can end up as evidence. This case won’t be the last time that happens.
- Context is everything. Asking a chatbot a hypothetical about cigarettes and fire reads very differently depending on who’s framing it. Prosecutors framed it as guilt. The jury saw curiosity.
- Juries are normalizing AI use. The juror’s “I talk to ChatGPT all the time” is the tell. When a behavior is common, it stops looking suspicious. That cuts against any case built on “he used a chatbot, therefore.”
What comes next
A mistrial isn’t an acquittal. Prosecutors can retry Rinderknecht, and if they do, you can expect them to lean harder on the physical evidence and treat the ChatGPT material as supporting color rather than a centerpiece. The lesson from the deadlock is clear enough: chat logs alone don’t carry a jury.
For everyone else, the takeaway is simpler and a little uncomfortable. The questions you type into an AI assistant aren’t sealed in a vault. They can be logged, requested, and read aloud in a courtroom. That doesn’t mean you should stop using these tools. It means treating them as private confessants is a mistake.
Expect more cases like this as AI chat becomes part of daily life and prosecutors test how far the evidence stretches. The Palisades trial suggests the public may be ahead of the courts on this one. For the full account, see the original report at The Verge AI.