Tag

Formal Verification

All articles tagged with #formal verification

Google DeepMind’s AI agents mathematically verify Signal’s encryption in four weeks
technology11 days ago

Google DeepMind’s AI agents mathematically verify Signal’s encryption in four weeks

Google DeepMind used its new AlphaProof Code AI framework to formally verify the security of Signal’s encryption protocol in just four weeks. The project proved that the post-quantum handshake and Triple Ratchet mechanisms function as intended at the machine-code level, marking a significant milestone for AI-assisted software security. While the verification did not uncover new vulnerabilities, it demonstrated that AI can drastically reduce the time and cost associated with traditional formal verification, which previously required years of human effort. The team emphasized that the process remains human-guided, with AI agents generating proofs that are checked against rigorous mathematical specifications.

Google DeepMind’s AlphaProof Code formally verifies Signal’s encryption in four weeks
technology12 days ago

Google DeepMind’s AlphaProof Code formally verifies Signal’s encryption in four weeks

Google DeepMind researchers used a new AI framework called AlphaProof Code to mathematically verify key security guarantees of the Signal protocol in just four weeks. The work validates the post-quantum handshake and Triple Ratchet but does not cover the entire app. While the project found no new vulnerabilities, it highlights the potential for AI to scale formal verification, a process previously too costly for widespread use.

AI Startups Bet on Automating Formal Verification to Secure Code
technology23 days ago

AI Startups Bet on Automating Formal Verification to Secure Code

A wave of startups is leveraging generative AI to automate formal verification, a mathematical process that proves software correctness. While traditional methods are too slow and costly for widespread use, new tools aim to make this rigorous testing practical for broader adoption, potentially reducing security vulnerabilities in critical systems.

Anthropic’s Claude Completes First Formal Proof of Fermat’s Last Theorem in 11 Days
technology2 months ago

Anthropic’s Claude Completes First Formal Proof of Fermat’s Last Theorem in 11 Days

Anthropic’s Claude AI generated the first fully computer-checked proof of Fermat’s Last Theorem in 11 days, producing 13 million lines of code. This automated effort outpaced a human-led project at Imperial College London that has been running since 2024. The proof, verified by mathematician Kevin Buzzard, resolves a 358-year-old problem by translating Andrew Wiles’s 1995 solution into a machine-verifiable format.

Terry Tao Leads an AI-Driven Reform of Mathematical Collaboration
science5 months ago

Terry Tao Leads an AI-Driven Reform of Mathematical Collaboration

Tao is promoting AI-assisted, computer-verified mathematics, shifting from solo work to massively collaborative, proof-checked research. Through Polymath-inspired projects and the Equational Theories initiative, he envisions solving thousands of subproblems with AI and Lean, creating a new, experimental era of mathematics where proofs are machine-checked for correctness.