The rise of AI chatbots like **OpenAI's ChatGPT**, **Anthropic's Claude**, and **Google's Gemini** has created an unprecedented digital landscape where users confide deeply personal information. These AI models, often by default, collect and store this sensitive data with few restrictions on its use, sharing, or potential disclosure to third parties or law enforcement. **Matt Green**, a privacy and security-focused computer science professor at **Johns Hopkins University**, highlights the profound implications: "You have this intelligent thing staring back at you, and you’re basically telling it, one question at a time, every possible thing there is to know about your life. You're giving it this huge profile on you." ### From Text Messages to AI: A Shifting Privacy Frontier **Moxie Marlinspike**, the cryptographer behind the end-to-end encrypted messenger **Signal**, points out a significant shift in the locus of privacy vulnerability. A decade ago, text messages represented the most sensitive data shared digitally. Now, he argues, AI interactions have taken that mantle, amplifying the surveillance risks. "Those same things I was concerned about with messaging are happening in the AI space, but several orders of magnitude more significantly," Marlinspike states. "People are integrating AI into their personal lives. They talk with it about their deepest insecurities, their finances, their health, their relationships." ### Confer: A New Standard for Private AI In response to these growing concerns, Marlinspike launched **Confer** earlier this year. This AI chatbot is designed to allow users to interact privately, employing cryptography to prevent the service's own servers from logging or surveilling conversations. As Marlinspike articulated in a blog post, "Confer is designed to be a service where you can explore ideas without your own thoughts potentially conspiring against you someday." Confer is part of an emerging wave of AI services that prioritize user privacy. While some promise not to record conversations or offer anonymization, a select few, like Confer, are implementing technological safeguards to limit their own access to user data. ### Understanding Zero Data Retention When engaging with major AI chatbots, users should assume a baseline of minimal privacy, especially if legal channels are pursued to access conversation records. This includes the service provider, advertisers, contractors involved in system fine-tuning, and law enforcement agencies. The most robust exception to this rule is **Zero Data Retention (ZDR)**. This contractual provision, primarily offered in enterprise versions of **OpenAI**, **Anthropic**, and **Google** AI services, legally obligates providers to immediately delete records of user interactions once processed. For privacy-conscious organizations and users, ZDR policies represent a critical safeguard in an increasingly data-hungry AI ecosystem.