TechFlow Logo
Login/ Sign up
ETH Gas
Gwei
Fear
gas
a16z Talks With a Stanford Professor: Can AI Really Break Cryptography? Unpacking the Truth Behind the Panic

a16z Talks With a Stanford Professor: Can AI Really Break Cryptography? Unpacking the Truth Behind the Panic

2026.10.12
Share
AI is becoming increasingly proficient at solving complex mathematical problems. What does this mean for the cryptography that protects private information, healthcare, banking, and crypto wallets?

Author: a16z crypto

Edited by: TechFlow

TechFlow Guide: The latest mathematical model breakthroughs from OpenAI have sparked industry panic, with some worrying that AI might dismantle the cryptographic systems we rely on. This article brings together a Stanford cryptography professor and the a16z crypto research team to separate fact from speculation, highlighting where AI actually presents opportunities. It is a must-read for anyone relying on cryptographic infrastructure.

Following warnings from industry insiders after OpenAI's mathematical breakthroughs, and concerns that AI might crack our foundational cryptography, we invited cryptographers from Stanford University and experts from a16z crypto to discuss what is fact, what is speculation, what warrants concern, and what does not.

Dan Boneh, Professor at Stanford University, heads the Applied Cryptography Group and co-leads the Computer Security Lab; also serves as a Special Research Advisor for a16z crypto

Eddy Lazzarin, General Partner at a16z crypto

Justin Thaler, Research Partner at a16z crypto, Associate Professor of Computer Science at Georgetown University

Liz Harkavy, Investment Partner at a16z crypto

Sonal Chokshi, Editor-in-Chief at a16z crypto

Additionally, six key takeaways:

1. AI's mathematical advances do not touch upon the core problems of cryptography

The recently released results demonstrate remarkable progress in mathematics and computer science, but they do not contain any cryptographic findings.

"This is a tremendous advance in mathematics and computer science, but there is absolutely nothing about cryptography in here... So you might think, does this mean these models are really bad at cryptography? We don't know.",Dan Boneh,03:00

"We didn't learn that elliptic curves are more fragile than before. We didn't learn that any specific cryptographic assumption is less secure than it was.",Eddy Lazzarin,15:19

2. Even if AI is smarter, it cannot find an algorithm that doesn't exist

A significant portion of modern cryptography rests on the assumption that no efficient solution exists for certain mathematical problems. This may hold true even for increasingly capable AI.

"We generally believe that no efficient algorithms exist for these problems. It's not that we think they might exist and humans just aren't smart enough to find them. We believe they simply don't exist. If they don't exist, it doesn't matter how smart the AI is. It can't find an algorithm that isn't there.",Justin Thaler,16:12

"We have to assume some problems are hard. We cannot prove that they are hard, precisely because that ties into a famous unsolved problem in computer science known as P versus NP. In a sense, we have to solve that first to truly argue that certain cryptographic problems are indeed, deeply hard... Whatever was published this week is nowhere near solving P versus NP. We still have to continually make assumptions in cryptography, right?",Dan Boneh,06:00

3. AI can compress years of cryptographic scrutiny

Attempting to break cryptography is itself one of the ways researchers build confidence. Only designs that withstand continuous, serious attacks earn trust. AI could change the pace at which this scrutiny occurs.

"In the past, when someone proposed an assumption and built a system around it, we'd often wait around 10 years to deploy the system to build confidence and give people enough time to analyze the assumption... And now, suddenly, we can use these models to drastically speed up this process... We can now run models for a month trying to break it, and if the model fails, our confidence in the validity of that assumption goes way up.",Dan Boneh,12:10

4. Quantum readiness provides a head start against some AI risks

While recent concerns focus on AI, some systems have already been targeted for migration due to the threat of quantum computing. This overlap has its limits.

"Due to concerns over quantum threats, people had already started viewing elliptic curve cryptography as something that might soon become vulnerable. Therefore, addressing these exact same concerns, we have actually already engaged, prepared, and deployed many of the mitigations you need, albeit originally for a different threat.",Eddy Lazzarin,20:52

"The threat from AI is broader in a sense because it could actually examine our post-quantum systems, weaken them in some way, or potentially break them entirely.",Dan Boneh,22:08

5. Cryptographic agility allows systems to adapt when weaknesses emerge

Cryptographic agility refers to a system's ability to swap out its cryptographic methods when evidence or requirements change. It equips engineers to handle discoveries they couldn't foresee in advance.

"The systems we deploy can quickly switch from one design to another if the one being used is proven insecure. Cryptographic agility is, in a sense, a friend to defending against surprises.",Dan Boneh,22:47

Dan points out that internet connections are already combining different cryptographic methods:

"They are already using agile cryptography. They are encrypting with both elliptic curves and lattices simultaneously. If you crack elliptic curves, it doesn't matter, because it is still encrypted under the lattice-based system.",Dan Boneh,38:40

6. The immediate opportunity: discovering more flaws before software release

One of the clearest opportunities for builders is to integrate more security testing throughout the development process.

When discussing complex cryptographic systems, Justin notes:

"The attack surface becomes extremely large, yes, and I am very optimistic that these AI tools that make people panic about cryptography will also help us clear out all the vulnerabilities. It's a very hard problem.",Justin Thaler,28:56

"Once a system is deployed, it attracts a lot of attention from cryptographers, but by then attacking the cryptographic system is actually too late. Now, because we control these AI models, any startup deploying new cryptographic capabilities, or even any company deploying any software, can have these models analyze vulnerabilities in its cryptography and overall software.",Dan Boneh,35:05

"Vulnerabilities can be eliminated. They can be designed around, they can be mitigated, and the primary resource we use to mitigate them, AI capability, is dropping dramatically in price every few months.",Eddy Lazzarin,37:10

Quotes have been slightly edited for clarity. Ellipses indicate omitted paragraphs.