Google DeepMind researchers used agentic AI to formally verify several key cryptographic guarantees of the Signal protocol, the widely used encryption system. Formal verification mathematically proves that software behaves as its designers intended. The researchers say they hope to extend the approach to all the software in the world.
What did Google DeepMind verify in Signal's encryption protocol?
Google DeepMind researchers used agentic AI to mathematically prove, or “formally verify,” a number of key cryptographic guarantees of the Signal protocol, the widely used encryption system at the center of the work.
The result matters for two fields at once. WIRED describes it as a landmark development for cryptography and for AI itself, pointing to the fact that the AI tools were used to confirm security properties that had to be checked rigorously rather than assumed.
What is formal verification?
Formal verification is a method of proving, with mathematical certainty, that a piece of software behaves the way its designers intended. In this case, the researchers were able to confirm that the widely used encryption protocol works as intended, a finding that carries weight for the millions of people who rely on it to keep their messages private.
The researchers are not stopping there. Their stretch goal is to apply the same approach to “all the software of the world,” a phrase that signals how far they hope the technique can reach if it scales.
The story was written by Lily Hay Newman, a senior writer at WIRED who covers information security, digital privacy and hacking. Her reporting is the primary account of the project available in the source material, and the details of the research methods and the specific guarantees verified are not described in the excerpt.
Dieser Artikel wurde mit Unterstützung von KI verfasst.
News Factory APP - agentische News für besseres SEO & AEO.