Google DeepMind just mathematically verified key parts of Signal
Summary
Google DeepMind researchers used its AlphaProof Code AI framework to formally verify several cryptographic guarantees in parts of the Signal protocol. The four-week study found no new vulnerabilities in the code it examined, but it did not verify the full Signal app, and the researchers say it remains unclear whether AI-assisted verification can scale affordably.Key Facts
- The analysis focused on key cryptographic guarantees in the Signal protocol, which is used by Signal, Google Messages and WhatsApp.
- DeepMind conducted the work as a four-week proof-of-concept project using AlphaProof Code.
- The researchers found no new vulnerabilities in the portions of the protocol they examined.
- When tested on older Signal versions with known, fixed vulnerabilities, the system identified them and showed the protocol could not be verified as written.
- The study did not formally verify the entire Signal mobile app.
- DeepMind is releasing the specification and proof it developed to support further formal verification work.
- The researchers had a head start from earlier formal verification work by the Signal Foundation.
- Google did not disclose the analysis’s cost or say whether it plans to offer a commercial formal verification product or service.
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
