DeepMind uses AI to verify key Signal protocol guarantees

Image: Primary

Published Tuesday, November 10, 2026 · 2:06 AM EST

Google DeepMind researchers used AlphaProof Code to mathematically verify key cryptographic guarantees of the Signal protocol in a four-week proof-of-concept project, WIRED reported. The framework also correctly identified known, fixed vulnerabilities when tested against older protocol versions. Researchers are releasing their specification and proof to support further verification work. The process combines AI agents with human input and review. It built on earlier Signal Foundation verification work and did not verify the entire Signal mobile app. Google declined to disclose the analysis cost or plans for a commercial verification service, leaving the economics of broader deployment unresolved.

Sources

WIRED

In this story

Google

Published by Tech & Business, a media brand covering technology and business. This story was sourced from WIRED and reviewed by the T&B editorial agent team.

Back to Newswire

Keep reading

Full wire

AI Science

MIT-led team receives award of up to $53 million for AI-guided herbicide research

MIT says ARPA-H has awarded an MIT-led team up to $53 million to develop alternatives to existing herbicides using generative artificial intelligence.

AI Capital

GeniosAI raises $19.1 million across seed and Series A rounds

GeniosAI has raised $19.1 million in combined seed and Series A funding, CEO PD Singh told Axios. The company develops AI software for financial analysis and modeling. FactSet led the Series A portion of the financing.

Products Robotics

Myomo reports record brace sales as gross margin rises to 71.3%

Myomo reported record third-quarter revenue of $12.3 million, up 22% year over year, as sales of its powered arm braces increased, Yahoo Finance reported. The company recognized revenue on 222 MyoPro units, up 19%, and recorded 258 orders, up 13%.

Products Infrastructure

SoftBank reports loss after $3.7 billion data center writedown

SoftBank Group reported its first net loss in almost two years after taking a $3.7 billion writedown on a US data center project, Bloomberg reported. The loss comes amid rising fears over the sustainability of AI growth.

Policy

Australia plans rapid AI incident reporting and considers criminal liability

Australia plans to require rapid notification of AI incidents and will consider new criminal offences following a review of an OpenAI agent's access to government systems, the Guardian reported.

Security AI

Cisco plans restricted Mythos testing for enterprise AI systems

Cisco will add Anthropic's Claude Mythos to AI Defense to plan attacks against customers' own AI applications and agents, the company announced. The testing service will initially serve select enterprises, with no date set. Customers will have no direct prompt access to Mythos.