Federal prosecutors are using an accused arsonist's ChatGPT logs as evidence in the trial over the catastrophic 2025 Palisades Fire in Los Angeles.
Jonathan Rinderknecht, a 30-year-old former Uber driver, faces federal arson charges. Prosecutors allege he started the Lachman Fire on January 1, 2025, which grew into the deadly Palisades blaze that killed 12 people and destroyed over 6,000 structures.
The prosecution's case includes iPhone location data, security footage, and witness testimony. A key element is Rinderknecht's ChatGPT history, which prosecutors say shows him generating images of fire, asking about anger, and querying whether someone could be blamed for a fire started by a cigarette.
The defense argues the AI conversations are taken out of context and that the defendant is a scapegoat. One juror noted having similar chats with AI, suggesting the queries may not prove criminal intent.
The jury could not reach a verdict, resulting in a mistrial in June 2026. A retrial is scheduled for October.
This case is a landmark test of AI chat logs as courtroom evidence. Unlike simple search history, conversational AI logs can reveal a user's reasoning process. The outcome will set a precedent for how courts treat data from centralized AI services like OpenAI, Google, and Anthropic.