{
  "id": 19328407,
  "title": "Introducing AlphaProof Code: a breakthrough in verifying complex software at depth",
  "url": "https://urgent.news/2026/11/10/introducing-alphaproof-code-a-breakthrough-in-verifying-complex",
  "topic": "tech",
  "section": "Tech",
  "published": "2026-11-10T12:02:28.000Z",
  "source": {
    "name": "Google DeepMind",
    "slug": "google-deepmind",
    "url": "https://deepmind.google/blog/introducing-alphaproof-code/"
  },
  "original_language": "en",
  "account": "Google DeepMind researchers have developed a groundbreaking framework, AlphaProof Code, which has successfully formally verified the cryptographic protocol behind Signal's privacy-preserving messaging system. This achievement marks the first time a widely deployed cryptographic protocol has been verified at such a depth, using AI agents in conjunction with human experts. The novel agentic framework, combining multiple AI models and a group of AI agents, streamlined the labor-intensive process of formal verification by translating Signal's security goals into formal mathematical specifications, generating proof through thousands of lines of Lean, and auditing the proofs for potential issues. The agents confirmed that the compiled code adheres to the critical security properties, such as safeguarding messages even when a device is temporarily compromised. This breakthrough not only demonstrates the potential of AI in enhancing software security but also opens the door to verifying other critical systems, including medical device firmware and power grid software. By providing mathematical proof of the protocol's security, this research enables businesses, governments, and critical infrastructure to bolster their defenses against modern cyber threats. The researchers have made their formal verification artifacts open-source, inviting the community to utilize and build upon this work. The collaboration between deep learning experts, security professionals, and industry leaders underscores the importance of integrating AI with traditional verification methods to create software that is mathematically proven to be secure.",
  "summary": "AlphaProof Code agents wrote three million lines of proofs in Lean to verify Signal’s core encryption.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}