Moxie Marlinspike, the creator of the privacy-focused messaging app Signal, is now setting his sights on transforming the realm of AI chatbots with a new initiative called Confer. This open-source AI assistant is designed to ensure that user data remains completely inaccessible to platform operators, hackers, law enforcement, and any unauthorized third parties. Confer employs strong encryption methods and operates on open-source software, allowing users to cryptographically verify its integrity. The platform uses a trusted execution environment (TEE) to encrypt conversations and data, ensuring that even server administrators cannot access or manipulate the information. Marlinspike's approach mirrors the simplicity and effectiveness that made Signal a popular choice for secure messaging. By removing the complexities often associated with encrypted communications, such as key management, Confer aims to provide a user-friendly experience while maintaining high privacy standards. This is particularly significant in light of the legal pressures that major platforms face, which often require them to hand over user data in response to subpoenas. Recent incidents, such as a court order compelling OpenAI to preserve all ChatGPT user logs, highlight the vulnerabilities of user privacy in AI platforms, where even sensitive interactions could be exposed. In contrast to existing AI services that may involve human oversight in monitoring conversations, Confer promises to uphold a stronger commitment to user privacy. By ensuring that conversations remain encrypted and securely stored on user devices, Marlinspike hopes to establish a new standard for privacy in AI interactions, akin to what he accomplished with Signal in the messaging space. This innovative endeavor reflects a growing need for tools that prioritize confidentiality in an era where data privacy concerns are increasingly prevalent.