Google DeepMind researchers used a new agentic AI framework, AlphaProof Code, in a four-week proof-of-concept to mathematically verify specific security guarantees of the Signal protocol. The team found no new flaws in the parts it examined. When tested on older versions with known issues, the AI identified the historical vulnerabilities.
The project shows how AI could help automate and speed up formal verification, a complex process for proving software security. It did not verify the entire Signal application. The Signal protocol also underpins encryption in WhatsApp and Google Messages. DeepMind is publishing its formal specification and proof for other researchers to build on.