Google DeepMind has unveiled AlphaProof Code, an AI system that mathematically proves software meets its security specifications. According to a blog post by DeepMind researchers Pushmeet Kohli and Swarat Chaudhuri, the system was demonstrated on Signal's encryption protocol.

The researchers say earlier verification efforts stopped at the theoretical design stage. AlphaProof Code instead verified the cryptographic core code that runs on users' devices.

The report says that after experts defined formal security requirements, AI agents built on Gemini Carbon wrote over six million lines in the Lean proof language. The agents flagged roughly one percent of the codebase for human review. Two bugs planted deliberately for testing were caught immediately.

The report states that the work took four weeks. It says verification of this kind typically takes years.

Signal CTO Ehren Kret says the approach strengthens trust in data privacy as AI threats grow. DeepMind plans to extend the method to power grids and medical devices through systems like CodeMender, according to the report.

Source: The Decoder