Earlier this year, we reported on the disturbing trend of AI chatbots actually helping individuals plan out violent attacks, and Florida prosecutors have already investigated ChatGPT for its purported role in a deadly shooting, but there’s another, even more famous incident that might have been planned out with the help of AI: the Pacific Palisades fire of 2025.

According to California prosecutors, suspected arsonist Jonathan Rinderknecht “used ChatGPT like a diary,” not only building up a fascination with fire but generating images of cities burning, reports the BBC. In one prompt pulled from Rinderknecht’s ChatGPT history, he asked the AI whether he could be held responsible for a fire caused by his fallen cigarette, according to ABC News. 

Rinderknecht was arrested in October of 2025 and charged with one count of destruction of property and one count of arson, with prosecutors seeking up to 45 years of prison time for his alleged role in sparking the fire that destroyed over 6,000 buildings and claimed 12 lives. 

“The evidence is strong that Jonathan Rinderknecht is responsible for igniting the fire on January 1, 2025, which eventually became the Palisades fire,” alleged US attorney Bill Essayli on an X social post. 

But when the trial of the United States v. Jonathan Rinderknecht concluded last Thursday, jurors were not persuaded. After two days of deliberation, the jurors informed the judge that they were unable to reach a decision, and a mistrial was declared. “There’s just not enough proof,” one juror told a CNN reporter. “A lot of holes.”

The drama is far from over, however. The judge has already scheduled a retrial for October of this year, with Rinderknecht to remain in custody until then.

Disclosure: Ziff Davis, Mashable’s parent company, in April 2025 filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.