🌐【海外AI最新情報】
Google DeepMindがAI「AlphaProof Code」でSignalの暗号プロトコルをコードレベルまで数学的に検証。毎日数十億件のメッセージを守る仕組みの安全性を、AIが初めて証明しました。
🔗元ツイート: xitter.twitterapi.xyz/GoogleDeepMind/status/…
#生成AI #AI最新情報 #海外AIニュース
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
16
SignalのコードをAIで数学的に証明した話、いちばん面白かったのは確かめ方。昔直したバグを2つわざと戻した版を用意して、ちゃんと「安全じゃない」と証明できるかまで試してる。
うちもAIの報告を別のAIに確かめさせてるけど、確かめ役にわざと間違いを混ぜて試すのは真似したい
#GoogleDeepMind
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
4
メッセージの「鍵」を、AIが数学で確かめた。Google DeepMindは、Signalの暗号の仕組みがコードの段階まで安全だと、AIで初めて数学的に証明したと発表した。「安全です」と言われて信じるか、証明を求めるか。毎日使うアプリの安全を、あなたは何で確かめているだろうか。
xitter.twitterapi.xyz/GoogleDeepMind/status/…
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
13
Protocolo de Signal: por primera vez una IA demostró matemáticamente que es seguro hasta el código. DeepMind lo verificó con AlphaProof Code; agentes de Gemini escribieron más de 6 millones de pruebas en Lean y los humanos solo revisaron el 1% crítico. xitter.twitterapi.xyz/GoogleDeepMind/status/…
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
9
值得注意的边界：这次证明的是 libsignal 一对一会话的核心代码（PQXDH 握手和三棘轮），而且假设底层密码原语本身可靠，并不等于整个 Signal App 都被证明安全。真正的门槛反而前移到人写的那约 2.6 万行规格上。想问一下：接下来会先把这套方法用在群聊，还是开放给其他开源加密库？
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
1
68
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
3
4. @GoogleDeepMind says AI has mathematically proven Signal's protocol secure down to the code level, using its AlphaProof Code framework. A big step for how software gets secured.
xitter.twitterapi.xyz/GoogleDeepMind/status/…
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
9
AI security budgets should shift toward specifying what must be true. A proof checker can reject bad reasoning; it cannot rescue a perfectly proved requirement that omits the failure you care about.
DeepMind’s AlphaProof Code produced roughly 6.5 million lines of Lean proofs for specified Signal security properties, resting on about 26,000 trusted lines of specifications and machine-code semantics. The economic opportunity is to automate proof construction while concentrating expert attention on those assumptions. That could let assurance scale without reviewing every generated line—but it does not make human judgment optional.
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
36
DeepMind just proved Signal's encryption down to the machine code.
6.5 million lines of Lean. Nobody will ever read them.
So I asked Claude Opus 5.5 to build a 3D film of what was actually proved.
xitter.twitterapi.xyz/GoogleDeepMind/status/…
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
12
W signal
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
2
131
Rule I'd steal from DeepMind's AlphaProof Code paper:
no prover agent, team lead or orchestrator can declare a goal done.
A gatekeeper does, once Lean checks the proof against a spec copy no agent can write to.
Who gets to say "done" in your setup? xitter.twitterapi.xyz/GoogleDeepMind/status/…
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
20
Demostrar que el código hace lo que dice la especificación. Para software crítico esto puede cambiar cómo se audita xitter.twitterapi.xyz/GoogleDeepMind/status/…
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
1
65
years of security verification compressed into four weeks, with more open proof work for other (ai) researchers to build on
@GoogleDeepMind Alphaproof code (an ai verification framework) proved specific security guarantees of the encryption code behind signal (the private messaging app). its agents generated over 6 million lines of lean proofs, with automated checks letting experts focus their review on roughly 1% of the codebase. in four weeks, they verified protections including keeping messages secure before a temporary device compromise and after recovery
also really important that deepmind is open-sourcing the verification artifacts, so researchers and ai agents can inspect, reproduce and build on the work
pretty exciting to me, beyond verifying signal’s security, useful parts could be reused to make future verification projects easier
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
33
Google DeepMind says AI has mathematically proven Signal's protocol is secure down to the code level.
It used its AlphaProof Code framework to verify the cryptographic system that protects billions of private messages every day.
xitter.twitterapi.xyz/GoogleDeepMind/status/…
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
9
As I always say: Use @signalapp!
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
44
6M+ lines of AI-written proofs, and you don't have to trust the AI: Lean checks them. What humans still own is the spec, and DeepMind says experts reviewed every one. A proof proves only what you asked it to. The new bottleneck is asking well.
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
5
AI just mathematically proved Signal's crypto is secure down to the code. so the only vulnerability left in the group chat is the friend who screenshots everything
@GoogleDeepMind did it with AlphaProof Code. which app do you want audited next? i'm saying my bank's
🤖 Made with AI
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
10
second, ai is starting to check work as well as produce it. google deepmind says ai has mathematically proven signal's protocol "is secure down to the code level," using its alphaproof code framework.
xitter.twitterapi.xyz/GoogleDeepMind/status/…
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
18
my guess is a year from now all crucial code, that millions of people use, will be formally verified and unhackable
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
1
7
Exactly the kind of good research and proof we need. nice one DeepMind team
For the first time, AI has mathematically proven @signalapp’s protocol is secure down to the code level.
Using our breakthrough AlphaProof Code framework, we verified the messaging app’s cryptographic system – which protects billions of private messages every day.
5
517