9to5Mac: DeepMind AI Reportedly Verifies Signal Protocol
Google DeepMind reportedly ran AI formal verification on the Signal protocol used by Signal, WhatsApp and Google Messages.
TL;DR
- 9to5Mac reports DeepMind used AI for formal verification of the Signal protocol.
- The four-week analysis flagged known weaknesses in older versions and built on prior non-AI work.
- It covers only the protocol, not full apps, and broader use faces heavy compute limits.
9to5Mac reports Google DeepMind ran an AI-driven formal verification on the Signal protocol that powers Signal, WhatsApp and Google Messages.
Formal verification means mathematically checking whether a protocol meets its security claims under every possible condition, a step beyond normal testing, according to 9to5Mac.
The outlet says the work is being called a landmark because it used agentic AI to hunt for any hidden failure points instead of just looking for known bugs.
9to5Mac notes the analysis took four weeks and built on earlier non-AI verification efforts. It adds that the system correctly flagged known weaknesses in older Signal versions.
DeepMind researcher Swarat Chaudhuri tells WIRED via 9to5Mac that scaling this to all software remains a long road.
The report stresses the check covers only the protocol itself, not full app implementations, and that broader use may stay limited by heavy compute needs.
The Tea
One outlet drops a big claim about AI finally stress-testing the encryption we all rely on. If it holds up, it could quiet some privacy doubts and push other teams to try the same trick. Still, four weeks and serious hardware for one protocol shows we’re not getting instant audits of everything tomorrow.