Signal

Google DeepMind used teams of Gemini Carbon agents to produce roughly 6.5 million lines of Lean proofs showing that Signal's actual compiled code keeps messages private, verified down to the processor instruction level and checked independently by Lean so no trust in the AI is required. The effort took about four weeks versus years for expert teams, with humans setting security goals and reviewing the ~1% flagged by agents. It is claimed as the first end-to-end formal verification of a widely used protocol at the machine-code level.

Google DeepMind introduced AlphaProof Code, a novel agentic framework for formal verification at unprecedented depth, and used it to guide AI agents in mathematically proving key security properties of the actual code behind the Signal protocol. The result is framed as a breakthrough on a long-standing grand challenge in software security, targeting real production code that protects billions of private conversations.