Google DeepMind just mathematically verified key parts of Signal
Summary
- Google DeepMind researchers used agentic AI via AlphaProof Code to mathematically prove key cryptographic guarantees of the Signal protocol.
- The four-week proof-of-concept project independently validated cryptography protecting platforms like Signal, Google Messages, and WhatsApp.
- When tested on older Signal versions containing known vulnerabilities, AlphaProof Code correctly identified them and failed verification.
- Researchers emphasize that while promising, scaling AI-automated formal verification to larger codebases remains a significant challenge.
Why it matters
AI-aided formal verification could revolutionize software security by mathematically proving code correctness and catching vulnerabilities, though scaling and cost questions remain.