Your AI chats are not as private as you think…

13 Sept 2026 · 1 min

Watch on YouTube

0:58

In this commentary, Dan Novaes warns that AI chat conversations, while feeling informal and temporary, are stored as digital data by the service providers. He explains that such stored data can be accessed in legal contexts, similar to emails or cloud documents, and therefore should not be assumed to be completely private. Novaes emphasizes the particular risk in workplace settings, where employees might inadvertently share confidential company information, customer data, or internal documents to AI tools. He advises users to treat AI inputs as data entered into someone else's system and to avoid sharing sensitive information.

Chapters

People tell AI things they wouldn't type into Google— work problems, financial questions, relationship issues, business plans, sometimes extremely personal information. But here's something many people don't think about: an AI conversation is still digital data. Depending on the service, those conversations may be stored for some period of time. And just like emails, text messages, or cloud documents, stored data can potentially become relevant during lawsuits or legal investigations. That doesn't mean someone can simply open your AI history whenever they want— legal processes and privacy protections still apply— but it does mean you shouldn't automatically treat an AI chatbot like a completely private conversation. That's especially important at work. Employees can accidentally paste confidential company information, customer data, or internal documents into AI tools. AI feels conversational, almost temporary, but the information you type into it can be much more permanent than the conversation feels. So before telling AI something sensitive, it's worth remembering you're still entering data into someone else's system.