Home · Technology · Nov 13 archive
Google DeepMind's AI Tool Validates Signal Protocol Security
Confirmed
In Short: Google DeepMind researchers announced they have used an AI tool called AlphaProof Code to produce a machine-checked mathematical proof that the encryption code behind Signal's messaging protocol behaves as designed.
AI Formal Verification of Signal Protocol | Google DeepMind, AlphaProof Code & Security | Uplatz. Uplatz
The proof covers Signal's post-quantum PQXDH handshake and Triple Ratchet, validating the design's forward secrecy, authentication, post-compromise security, and handshake secrecy.
In a test, the team reintroduced two bugs that had previously been found and fixed in Signal's code, demonstrating the tool's ability to identify these issues.
Signal's chief technology officer, Ehren Kret, said, 'Using AI agents to translate live code into valid Lean theorems is a major breakthrough.'
The researchers translated Signal’s security goals into formal mathematical specifications, marking the first time AI has been employed to verify a widely deployed cryptographic protocol at such depth.
Google DeepMind ran AlphaProof Code against the Signal protocol for four weeks and found no new flaws in the parts they examined.
The framework also identified known weaknesses in older versions of Signal containing vulnerabilities, confirming its robustness.
The work demonstrates Google DeepMind's formal verification capabilities using a new AI framework called AlphaProof Code.
Google DeepMind researchers used agentic AI to mathematically prove several key cryptographic guarantees of the Signal protocol, the encryption system behind Signal, Google Messages, and WhatsApp.
The researchers hope to apply this technology during the software design process, not just to cryptographic protocols but to code in general.
Pushmeet Kohli, chief scientist at Google Cloud and DeepMind's vice president of research, said, 'If you can now verify the correctness of these implementations, then you can rest easy.'
What this adds
This marks the first time AI has been employed to verify a widely deployed cryptographic protocol at such depth, addressing a long-standing challenge in cybersecurity.
The framework demonstrated its robustness by successfully identifying insecurities in isolated versions of Signal’s code containing reintroduced, previously fixed bugs.
Google DeepMind is releasing the specification and proof it built for the Signal analysis, in the hope of helping other formal verification projects.
Background
Google DeepMind researchers announced they have used an AI tool called AlphaProof Code to produce a machine-checked mathematical proof that the encryption code behind Signal's messaging protocol behaves as designed.
What's confirmed
- "If you can now verify the correctness of these implementations, then you can rest easy," said Pushmeet Kohli, chief scientist at Google Cloud and DeepMind's vice president of research.
- "It's a system that is fundamentally built for human AI collaboration," Swarat Chaudhuri, a research scientist at Google DeepMind and a professor at the University of Texas at Austin, said in an interview.
- "While prior formal verification focused on smaller portions of the code, Google DeepMind's approach enables significant expansion," Signal chief technology officer Ehren Kret said in a statement.
- In a test of the system, the team reintroduced two bugs that had previously been found and fixed in Signal's code.
- In this case the proof covered Signal's post-quantum PQXDH handshake and Triple Ratchet, validating the design's forward secrecy, authentication, post-compromise security and handshake secrecy.
- The process began with researchers and AI agents translating Signal’s security goals into formal mathematical specifications.
- Researchers at Google DeepMind ran AlphaProof Code against the Signal protocol for four weeks and found no new flaws in the parts they examined.
- And it demonstrates Google DeepMind's formal verification capabilities using a new AI framework called AlphaProof Code.
- Google DeepMind researchers used agentic AI to mathematically prove, or “formally verify,” a number of the Signal protocol’s key cryptographic guarantees—a landmark development for the widely used cryptography and AI itself.
- Photo: Helar Lukats, Wikimedia Commons, CC BY-SA 4.0 8 hours ago 8 hours ago WORK Google DeepMind’s AI wrote 6.5 million lines of proof for Signal’s encryption in four weeks Google DeepMind says its AI agents formally verified the core of Signal's encryption, the protocol under WhatsApp, in about four weeks.
- Pointing to a related DeepMind project, AlphaProof Nexus, which is focused on mathematics, Chaudhuri said the next step is to apply this technology during the software design process – not just to cryptographic protocols but code in general.
What's still developing
- The work, announced in a DeepMind blog post on Monday, marks the first time a widely deployed cryptographic protocol has been formally verified down to the exact code running on people’s devices, by AI or by any other method.
- Signal president Meredith Whittaker issued a blunt warning on June 20 that AI agents like Microsoft Copilot are surveillance infrastructure in disguise, arguing their sweeping access to contacts, messages, and credentials effectively nullifies end-to-end encryption.
- Google's AI wrote checked proofs for Signal's code.
- Source : Google DeepMind melds humans and AI agents to vet Signal's secure messaging techniques, by Thomas Claburn, theregister, Nov 10, 2026 Adobe announced always-on AI agents for Firefly at its Max conference in Miami Beach, with a beta planned for early 2027.
- In so doing, the Google/DeepMind team – Swarat Chaudhuri, Andrew Ferraiuolo, Anton Kovsharov, Tej Chajed, Sam Schlesinger, Karn Seth, Jonathan Katz, Charles Sutton, and Pushmeet Kohli – has taken a step toward a challenge the Beneficial AI Foundation issued in April to formally verify the Signal protocol and Signal app with Lean.
- However, encouragement can be taken from the fact that when tested against older versions of Signal that contained known vulnerabilities, it identified those and flagged them as failing formal verification.
- Ars Technica has been separating the signal from the noise for over 25 years.
- The proofs cover the parts of Signal that guard a conversation: the PQXDH handshake that sets up a chat and the Triple Ratchet that changes the key for every message, including the post-quantum layer meant to resist future quantum computers.
- Since 2023, Signal has been upgrading its encryption technology so that it remains secure against the future, hypothetical threat of quantum computers.
- Google DeepMind said the technology could make advanced software security checks faster and less expensive, potentially extending them to critical systems such as medical devices and power grids.
- Agentic AI can significantly cut down the time of the latter project, enabling analyses like DeepMind's Signal evaluation.