Moxie Marlinspike, the founder of Signal Messenger, has unveiled Confer, a new AI chatbot designed to bring the same level of privacy to artificial intelligence that Signal brought to messaging.
Unlike mainstream platforms such as ChatGPT or Gemini, Confer encrypts every conversation end-to-end, ensuring that even server administrators cannot read user data.
How Confer protects your chats
Confer uses end-to-end encryption for all interactions, meaning only the account holder can access their messages.
Instead of relying on passwords, the system generates encryption keys through Face ID or device PIN. It also runs inside a trusted execution environment (TEE), which keeps data encrypted even while responses are processed.
This approach prevents leaks, hacks, or misuse of personal information.
The platform is built entirely on open source software, allowing users to cryptographically verify that the privacy protections are in place.
This transparency echoes Signal’s philosophy of trust through verifiability. In making its large language models and back-end components open, Confer aims to reassure users that their conversations remain private, even from law enforcement or hackers.
Why it counts
AI chatbots have exploded in popularity, but concerns about data privacy remain.
Many platforms store conversations for training or advertising, which undoubtedly raises questions about how personal information is used.
Confer positions itself as a consumer-friendly alternative for those who value privacy above convenience. If you trusted Signal to keep your messages secure, Confer offers a familiar promise: your words indeed stay yours.
Confer’s launch highlights a growing demand for ethical AI tools that prioritise user rights.
While competitors are racing to add features, Confer builds on trust and transparency. It may not replace mainstream chatbots overnight, but it sets a new benchmark for what private AI should look like.
