Palisades Fire suspect's ChatGPT history to be used as evidence (www.rollingstone.com)

The Department of Justice announced the arrest of 29-year-old Jonathan Rinderknecht in connection with the January Palisades Fire — a blaze that killed 12 people and destroyed thousands of structures. He’s federally charged with “destruction of property by means of fire,” a felony carrying a five‑year mandatory minimum. Prosecutors highlighted “digital evidence” in their affidavit: months before the wildfire, Rinderknecht allegedly asked ChatGPT to generate iterations of a dystopian image showing a burning forest and fleeing crowds, and those images were shown at a press conference alongside phone records, alleged false statements, and video of his behavior near an earlier brush fire investigators say he started. For the AI/ML community the case is notable because it signals how chatbot logs and generated content can be treated as forensic evidence. OpenAI’s policy says content or user data is disclosed only in response to valid legal process, and CEO Sam Altman has warned there’s no legal confidentiality with ChatGPT — yet it’s unclear if OpenAI cooperated here. Courts are increasingly treating AI interaction records like other discoverable digital communications, raising implications for model logging, data retention, user privacy, and compliance workflows. Developers, operators and users should expect chatbot histories (including text prompts and generated images) to be subject to subpoenas or warrants, and consider technical and policy safeguards around access controls, retention limits, and disclosure procedures.