Let's Encrypt AI wants every AI chat protected by confidential computing by 2027

Let's Encrypt AI: protect AI chats with on AI GPUs by the end of 2027

An independent position paper argues that prompts, responses, and context should stay private during AI inference, using hardware-backed trusted execution environments and remote attestation on GPUs. It lays out a standard: encrypt the whole route, make security-critical software inspectable, keep content under user control, and make verification a condition of service. The goal is to make confidential computing the default on most AI GPUs by the end of 2027, starting with open-weight models and using buyer demand to push frontier labs to offer it.