DeepMind's AlphaProof Code proves security properties of Signal's libsignal code

TL;DR

  • On 10 November Google DeepMind introduced AlphaProof Code, a framework in which people decide what to prove about a program and AI agents write the proofs.
  • DeepMind used it to prove key security properties of libsignal, the Signal protocol's reference implementation, at the machine-code level: about 6.5 million lines of Lean proof on about 26,000 trusted lines, the paper says.
  • The proofs cover a command-line build of the library's one-to-one session code, and the paper says the results don't concern Signal's own applications.

Read the full story

Sign in with your email to read AI News. It’s free.

Share this story

Explain like I’m 15