Your private conversations with ChatGPT aren't private.

Prosecutors just used ChatGPT logs as evidence to help convict Jonathan Rinderknecht for arson in the deadly Palisades fire case. This isn't some distant legal theory anymore. Your AI chat history is fair game in court.

Most people treat ChatGPT like a personal assistant or therapist. They ask sensitive questions, share personal details, work through problems. The assumption? These conversations stay between you and the AI.

Wrong.

What Happened in the Palisades Case

Rinderknecht faced arson charges for starting a fire that became one of LA's deadliest wildfires. Prosecutors built their case with iPhone location data, security footage, and witness testimony.

But they also pulled his ChatGPT conversation logs.

We don't know exactly what those conversations contained. The details haven't been released. But the fact that prosecutors used them suggests they found something incriminating. Maybe questions about fire behavior. Maybe searches for remote locations. Maybe something that showed premeditation.

The point is: they looked, they found something useful, and a judge allowed it as evidence.

Your AI Conversations Aren't Protected

Here's what most people don't realize: ChatGPT conversations aren't covered by attorney-client privilege. They're not protected like conversations with your doctor or therapist. They're just data sitting on OpenAI's servers.

OpenAI's privacy policy is clear. They can share your data when legally required. That includes subpoenas, court orders, and law enforcement requests.

Think about what you've asked ChatGPT lately:

  • Help writing breakup texts
  • Advice on workplace conflicts
  • Questions about financial problems
  • Research on sensitive topics
  • Personal struggles and fears

All of that could end up in court if you're ever involved in a legal case. Divorce proceedings. Criminal charges. Civil lawsuits. Employment disputes.

The Bigger Privacy Problem

This goes beyond ChatGPT. Every AI service you use creates a digital trail:

  • Claude conversations through Anthropic
  • Copilot chats through Microsoft
  • Gemini interactions through Google
  • Local AI models that phone home

Some services offer "incognito" or "temporary" chat modes. But even then, there's no guarantee. Companies can change policies. Governments can demand access. Hackers can breach systems.

The assumption of privacy was always an illusion. Now we have proof.

What You Can Do Right Now

First, audit your AI chat history. Most services let you download or delete your conversation data:

  • OpenAI: Go to Settings > Data Controls > Export Data
  • Delete conversations you wouldn't want in court
  • Use temporary/incognito modes for sensitive topics

Second, change how you interact with AI. Treat it like talking to a stranger in public, not a private therapist. Ask generic questions instead of personal ones. Don't include identifying details in your prompts.

Third, consider local AI models for truly sensitive work. Tools like Ollama let you run AI on your own computer. No cloud servers. No conversation logs. But remember: these models are less capable than commercial options.

The New Reality

This case sets a precedent. Expect more prosecutors to request AI conversation logs. Expect more judges to allow them as evidence. Expect more privacy erosion disguised as technological progress.

The companies won't save you. They'll comply with legal requests every time. Their business model depends on government cooperation, not your privacy.

Your AI assistant knows everything you've told it. Soon, everyone else might too.

— Dolce