When Jonathan Rinderknecht faced arson charges for setting a fire on New Year’s Day 2025 — one that became among the deadliest wildfires in LA history — prosecutors pulled out every tool they had. Location data from his iPhone. Security camera footage. Witness testimony. And, notably, his ChatGPT conversation logs.
The prosecution argued the logs revealed intent. Rinderknecht had asked ChatGPT to generate images of fire, asked the chatbot “Why am I so angry all the time?”, and ranted about the wealthy destroying the world. In one screen recording, he asked ChatGPT whether someone could be blamed for a fire lit by their cigarette.
The jury wasn’t convinced. The trial ended in a deadlock — 10-2 in favor of the defense. The judge declared a hung jury and a mistrial.
One juror later said she didn’t believe ChatGPT logs proved anything. “I talk to ChatGPT all the time,” she told CBS LA. She said it actually made her angry that they were treating his chatbot use as a character flaw.
This case raises a question more courts will soon face: do your AI conversations constitute admissible evidence? And if so, what does it mean when half the country uses these tools daily for everything from brainstorming to venting? The jury’s reaction suggests that at least for now, chatbot logs alone aren’t enough to convict.
