Your ChatGPT conversations are not as private as they seem – they can be revealed in court. Unlike communications with licensed professionals like psychologists, doctors, or lawyers, interactions with ChatGPT do not enjoy the same level of legal protection and confidentiality.
This was stated by OpenAI CEO Sam Altman during a podcast with Theo Von on YouTube, reports Baltimore Chronicle, with reference to his comments.
Altman emphasized that, in the event of a legal case, the company might be required to provide a transcript of a user’s conversation with the chatbot at the court’s request. Therefore, as he pointed out, if you share personal or sensitive information with AI, you should be aware that these conversations could potentially become available to others. Unlike encrypted messaging apps like Signal or WhatsApp, conversations with ChatGPT can be reviewed by the company, as OpenAI partially uses these data to improve its models and detect potential policy violations.
While the company claims that user messages in both the free and paid versions of ChatGPT are retained for no more than 30 days, in some cases, they may be stored for longer periods due to security reasons or legal requirements. This raises concerns about privacy, especially as people’s trust in AI for solving personal issues continues to grow. Altman believes that communication with language models should be protected in the same way as consultations with professionals, urging swift legislative action to address this issue.
Meanwhile, OpenAI is preparing to launch the new, more powerful GPT-5 model, which is expected to offer conversations nearly indistinguishable from those with a human.
OpenAI continues to face competition from Elon Musk, who has unveiled Grok 4 – an AI system that he claims is “the smartest AI ever created.” However, it has been revealed that the chatbot frequently draws answers from Musk’s personal posts, leaving users puzzled.
Earlier we wrote that Musk raises $6 billion to boost xAI as OpenAI rival.