Beneficial AI Foundation

graninas

ByteDance

ByteDance Seed Team

AI Magazine

Why Is Anthropic Not Releasing Claude Mythos To The Public

Anthropic

Claude Mythos Preview \ red.anthropic.com

Tom's Hardware

Anthropic's latest AI model identifies 'thousands of zero-day vulnerabilities' in 'every major operating system and every major web browser'  Claude Mythos Preview sparks race to fix critical bugs, s

Lean

Lean Programming Language

The Heartbleed Bug

Heartbleed Bug

Metaculus

When will AIs program programs that can program AIs?

Lesswrong

Limitations on Formal Verification for AI Safety  LessWrong

ABC News

What to know about Signal, which the Pentagon previously discouraged workers from using

Business of Apps

Signal Revenue & Usage Statistics (2026)

The Heartbleed security flaw that affects most of the Internet

FOX News

What you need to know about the Heartbleed bug

seL4

seL4 in use

Association for Computing Machinery

1629575.1629596

Formally Verified Software In The Real World

signal

Signal Protocol and Post-Quantum Ratchets

WhatsApp's Signal Protocol integration is now complete

Nature

Olympiad-level formal mathematical reasoning with reinforcement learning

Elon Musk

Rishi Sunak

Kamala Harris

Bill Gates

Andrew Ng

Yoshua Bengio

Fei-Fei Li

Demis Hassabis

Melanie Mitchell

Sam Altman

Dario Amodei

Geoffrey Hinton

Yann LeCun

Gary Marcus

Max Tegmark

Connor Leahy

Eliezer Yudkowsky

Jaan Tallinn

Marc Andreessen

Eric Schmidt

Norbert Wiener

Arthur Clarke

Irving John Good

Claude Shannon

Hans Moravec

John Smart

There's a "wall of fear-mongering and doomerism" in the AI world right now.

"Worrying about AI today is like worrying about overpopulation on Mars."

Concerns that AI could pose a threat to humanity is "preposterously ridiculous."

Claiming AI poses an existential threat is "such an extreme" and risks "wip[ing] out some of its potential benefits."

"I'm more concerned about... the risks that are here and now [than the existential threat of AI]."

Current AI is "not anywhere close" to posing an existential threat but it could in the future.

AI is the "biggest existential threat" to humanity.

AGI's worst-case scenario would be "lights-out for all of us."

Powerful AI systems "taking control" pose an "existential threat."

AI has a "10 to 25 per cent" chance of destroying humanity.

"The most likely result of building a superhumanly smart AI... is that literally everyone on Earth will die."

Does AI Pose an Existential Threat to Humanity?

The Beneficial AI Foundation (BAIF), co-founded by MIT professor Max Tegmark and Tufts Researcher Meia Chita-Tegmark, on Monday launched Signal Shot, a formal verification project using Lean theorem prover and AI tools to produce a mathematical proof that Signal's encryption protocol is secure in both design and real-world code.

The Beneficial AI Foundation (BAIF), co-founded by MIT professor Max Tegmark and Tufts Researcher Meia Chita-Tegmark, on Monday launched Signal Shot, a formal verification project using Lean theorem prover and AI tools to produce a mathematical proof that Signal's encryption protocol is secure in both design and real-world code.

Formal verification uses mathematical proof to confirm a system cannot behave outside its specification. It has previously been applied in aerospace and autonomous aviation, with the seL4 microkernel described as the first formally verified general-purpose operating system kernel. BAIF claims Signal Shot would represent its first application to a messaging protocol of this scale.

Formal verification uses mathematical proof to confirm a system cannot behave outside its specification. It has previously been applied in aerospace and autonomous aviation, with the seL4 microkernel described as the first formally verified general-purpose operating system kernel. BAIF claims Signal Shot would represent its first application to a messaging protocol of this scale.

BAIF states that Signal Shot will verify security by "writ[ing] down a mathematical model of the protocol and prov[ing] its security properties," before translating Signal's actual Rust code into Lean via a tool called Aeneas and proving it matches that model — producing a logical proof rather than a test result.

BAIF states that Signal Shot will verify security by "writ[ing] down a mathematical model of the protocol and prov[ing] its security properties," before translating Signal's actual Rust code into Lean via a tool called Aeneas and proving it matches that model — producing a logical proof rather than a test result.

The initiative is organized by BAIF in partnership with Lean-focused group LeanFRO and in coordination with Signal. AI will assist by "auto-formalizing research papers, expanding proof sketches into detailed verification plans" and "grinding through the tedious proof steps," work it describes as having previously been "bottlenecked by human effort."

The initiative is organized by BAIF in partnership with Lean-focused group LeanFRO and in coordination with Signal. AI will assist by "auto-formalizing research papers, expanding proof sketches into detailed verification plans" and "grinding through the tedious proof steps," work it describes as having previously been "bottlenecked by human effort."

The effort builds on recent advances in AI-assisted reasoning, including the Claude Mythos system from Anthropic, reportedly highly effective at identifying software vulnerabilities but not publicly released due to concerns it could be misused to exploit critical infrastructure. The Lean Theorem Prover has also been central to recent breakthroughs in LLM-based mathematical reasoning, including gold medal-level performances at the 2025 International Math Olympiad.

The effort builds on recent advances in AI-assisted reasoning, including the Claude Mythos system from Anthropic, reportedly highly effective at identifying software vulnerabilities but not publicly released due to concerns it could be misused to exploit critical infrastructure. The Lean Theorem Prover has also been central to recent breakthroughs in LLM-based mathematical reasoning, including gold medal-level performances at the 2025 International Math Olympiad.

Signal’s protocol, which underpins end-to-end encryption for over one billion users via its integration into WhatsApp in 2016, has previously incorporated formal verification into its design through work with Cryspen on the Triple Ratchet protocol. However, past failures such as the Heartbleed Bug in 2014 have exposed how gaps between verified design and implementation can persist, after a flaw in OpenSSL went undetected for years and compromised widely used cryptographic systems.

Signal’s protocol, which underpins end-to-end encryption for over one billion users via its integration into WhatsApp in 2016, has previously incorporated formal verification into its design through work with Cryspen on the Triple Ratchet protocol. However, past failures such as the Heartbleed Bug in 2014 have exposed how gaps between verified design and implementation can persist, after a flaw in OpenSSL went undetected for years and compromised widely used cryptographic systems.

Beneficial AI Foundation #

ByteDance #

AI Magazine #

Anthropic #

Tom's Hardware #

Lean #

The Heartbleed Bug #

Metaculus #

Lesswrong #

ABC News #

Business of Apps #

CNN #

FOX News #

seL4 #

Association for Computing Machinery #

signal #

Nature #

AI will handle tasks including "auto-formalizing research papers, expanding proof sketches into detailed verification plans" and "grinding through the tedious proof steps," work it describes as having previously been "bottlenecked by human effort."

AI will handle tasks including "auto-formalizing research papers, expanding proof sketches into detailed verification plans" and "grinding through the tedious proof steps," work it describes as having previously been "bottlenecked by human effort."

Signal has previously integrated formal verification into its protocol design, partnering with Cryspen on the Triple Ratchet protocol. According to Business of Apps, the platform had 70 million active users as of 2024. However, over one billion monthly active users were brought under its encryption in 2016 when WhatsApp completed its integration of the Signal Protocol.

Signal has previously integrated formal verification into its protocol design, partnering with Cryspen on the Triple Ratchet protocol. According to Business of Apps, the platform had 70 million active users as of 2024. However, over one billion monthly active users were brought under its encryption in 2016 when WhatsApp completed its integration of the Signal Protocol.

In 2014, a bug known as The Heartbleed Bug arose between a verified protocol design and its implementation in OpenSSL, a cryptographic software library used by an estimated two-thirds of web servers. It passed quality assurance tests, went undetected for over two years, and ultimately exposed private keys to attackers.

In 2014, a bug known as The Heartbleed Bug arose between a verified protocol design and its implementation in OpenSSL, a cryptographic software library used by an estimated two-thirds of web servers. It passed quality assurance tests, went undetected for over two years, and ultimately exposed private keys to attackers.

For decades, software security has meant manually finding bugs faster than attackers do. AI and formal methods, as embodied by BAIF's Signal Shot, may finally break that cycle. If machines can now handle the proof work that once took decades of human effort, cryptographic guarantees stop being a research curiosity and become a fundamental breakthrough transforming encryption security forever. Signal Shot brings that future of provably secure communication within reach.

For decades, software security has meant manually finding bugs faster than attackers do. AI and formal methods, as embodied by BAIF's Signal Shot, may finally break that cycle. If machines can now handle the proof work that once took decades of human effort, cryptographic guarantees stop being a research curiosity and become a fundamental breakthrough transforming encryption security forever. Signal Shot brings that future of provably secure communication within reach.

No proof, however rigorous, can fully capture the complexity of constantly-updating, real-world code running on real-world hardware. Heartbleed was a single rogue variable in an implementation nobody had formally modeled, rather than an inherent design flaw. Signal runs on millions of different devices, operating systems, and network conditions, making Signal Shot a commendable but near-impossible mission.

No proof, however rigorous, can fully capture the complexity of constantly-updating, real-world code running on real-world hardware. Heartbleed was a single rogue variable in an implementation nobody had formally modeled, rather than an inherent design flaw. Signal runs on millions of different devices, operating systems, and network conditions, making Signal Shot a commendable but near-impossible mission.

There is a 50% chance that AIs will program programs that can program AIs by April 2028, according to the Metaculus prediction community.

There is a 50% chance that AIs will program programs that can program AIs by April 2028, according to the Metaculus prediction community.

The AI Doc: Or How I Became an Apocaloptimist

The Singularity Is Nearer

If Anyone Builds It, Everyone Dies

Artificial Intelligence

Matthias Balk/Picture Alliance/Getty Images

BAIF Launches 'Signal Shot' to Verify Signal Protocol Using AI

Science & technology

World

Communication

Computer security

For decades, software security has meant chasing bugs faster than attackers. By combining AI with tools like Lean, Signal Shot aims to shift that paradigm toward mathematical certainty. If machines can scale formal proofs once limited by human effort, provable guarantees could become practical, bringing the long-promised vision of truly secure communication closer to reality.

For decades, software security has meant chasing bugs faster than attackers. By combining AI with tools like Lean, Signal Shot aims to shift that paradigm toward mathematical certainty. If machines can scale formal proofs once limited by human effort, provable guarantees could become practical, bringing the long-promised vision of truly secure communication closer to reality.

Even the most rigorous proof cannot fully capture messy real-world systems. Signal runs across diverse devices, codebases, and updates, where unmodeled edge cases can emerge. Projects like Signal Shot may strengthen assurance, but cases like Heartbleed suggest full, end-to-end proof of real-world security remains an elusive, and perhaps even unattainable, goal.

Even the most rigorous proof cannot fully capture messy real-world systems. Signal runs across diverse devices, codebases, and updates, where unmodeled edge cases can emerge. Projects like Signal Shot may strengthen assurance, but cases like Heartbleed suggest full, end-to-end proof of real-world security remains an elusive, and perhaps even unattainable, goal.

BAIF Launches 'Signal Shot' to Verify Signal Protocol Using AI

BAIF Launches 'Signal Shot' to Verify Signal Protocol Using AI

The Spin

Metaculus Prediction

The Controversies

Andreessen

Ng

LeCun

Mitchell

Li

Bengio

Musk

Altman

Hinton

Amodei

Yudkowsky

Go Deeper

Sign Up for Our Free Newsletters
Sign Up for Our Free Newsletters

BAIF Launches 'Signal Shot' to Verify Signal Protocol Using AI

BAIF Launches 'Signal Shot' to Verify Signal Protocol Using AI

The Spin

Metaculus Prediction

The Controversies

Andreessen

Ng

LeCun

Mitchell

Li

Bengio

Musk

Altman

Hinton

Amodei

Yudkowsky

Go Deeper

Sign Up for Our Free NewslettersSign Up for Our Free Newsletters

Sign Up for Our Free Newsletters
Sign Up for Our Free Newsletters