AI Chats as Court Evidence: The Case That Sets a Precedent

What happened?

In January 2025, Jonathan Rinderknecht was accused of starting one of the deadliest wildfires in Los Angeles history, the Palisades Fire, which scorched over 23,000 acres and destroyed hundreds of homes. To build their case, prosecutors turned to an unprecedented element: records of his conversations with ChatGPT. According to The Verge, the evidence includes prompts from the defendant asking the chatbot to generate images of fire, questions like "Why am I so angry all the time?" and a screen recording where Rinderknecht asks if someone can be blamed for a fire if it was started by their... The prosecution also used location data from his iPhone, security camera footage, and witness testimony. This case marks a milestone in the use of generative artificial intelligence as evidence in a criminal trial.

Why is it important?

This case marks the first time records from a conversational AI assistant have been admitted as evidence in a criminal trial. It raises fundamental questions about privacy in AI interactions, the scope of data protection, and the right against self-incrimination. Unlike emails or text messages, AI chats can contain intimate reflections, unfiltered thoughts, and even simulations of hypothetical scenarios that could be misinterpreted out of context. Rinderknecht's defense has requested a mistrial, arguing that the evidence obtained from ChatGPT violates his right to privacy and that the content may have been manipulated or misinterpreted.

Legal implications

• Right to privacy: To what extent are conversations with AI protected by the Fourth Amendment (in the US) or regulations like GDPR in Europe? In the United States, the third-party doctrine holds that information shared with a third party is not protected, but AI is not a traditional third party; it is a service that learns from data.
• Probative value: Judges will need to determine whether these conversations are reliable, authentic, and not manipulable. In this case, the prosecution presented a screen recording, but the defense questions whether the content could have been edited or whether the chatbot generated responses that do not reflect real intentions.
• Precedent: Other courts could follow this example, opening the door for prosecutors and defense attorneys to request records from Alexa, Siri, Google Assistant, and other assistants. This could lead to a flood of data requests to tech companies.

What consequences will it have?

In the short term, the case could end in a mistrial if the defense successfully argues that the evidence was obtained illegally or that its use violates constitutional rights. In the long term, we will see an increase in requests for data from AI companies by authorities, as well as possible changes in these companies' privacy policies. Specific judicial guidelines for the admissibility of this type of evidence are also likely to emerge. AI companies like OpenAI, which develops ChatGPT, face growing pressure to balance service utility with privacy protection. In response, they could implement end-to-end encryption features or allow users to automatically delete their histories.

"AI is not a private diary, but it is not a neutral witness either. Courts will have to decide where to draw the line," says a digital law expert consulted by TheVortiq.

What should readers know?

If you use AI assistants, you should be aware that your conversations could be used against you in legal proceedings. AI companies typically store chat histories to improve their models, and although they claim to anonymize data, authorities can request user identity through court orders. In the case of ChatGPT, OpenAI retains conversation records for a period that may vary, and users can opt out of contributing to model training, but conversations are still stored. We recommend reviewing each service's privacy policies and, if you are part of an investigation, consulting with a lawyer before deleting or modifying any records, as doing so could be considered obstruction of justice.

Comparison with previous cases

This is not the first case where personal digital data has been used as evidence. In 2022, prosecutors used data from a smart thermostat to accuse a man of murder. In 2023, in a Florida murder case, recordings from an Amazon Echo smart speaker were used. However, the conversational and generative nature of AI adds a layer of complexity: the content is not a mere record of facts, but an interaction where the user may be exploring ideas, asking hypothetical questions, or simply venting. In this case, the defense could argue that Rinderknecht's questions to ChatGPT were speculative or cathartic, not evidence of criminal intent. Moreover, unlike a personal diary, conversations with AI are partly generated by the model itself, raising questions about authorship and reliability.

Conclusion

The Rinderknecht case is a milestone at the intersection of artificial intelligence and criminal law. As technology advances, legal systems will need to adapt to balance the pursuit of justice with the protection of fundamental rights. The judge's decision on the mistrial motion will be crucial in defining the future of privacy in AI interactions. We will be watching the evolution of this trial and any future regulations that may emerge, both in the United States and other jurisdictions, where the European GDPR could offer a stricter framework for personal data protection.