AI Could Aid Attacks on Cryptography Before Quantum Computers Arrive

Zcash co-founder and StarkWare CEO Eli Ben-Sasson warned Thursday, Nov. 5, that artificial intelligence (AI) could help attackers target cryptography before quantum computers arrive, but identified no specific attack or affected crypto network.

In a Nov. 5 post, Ben-Sasson called for systems to move toward what he described as the strongest cryptography, writing: “Systems need to migrate to the strongest forms of cryptography: hash-based cryptography (like ZK-STARKs).” He did not lay out a migration plan or explain how developers should apply the recommendation to wallets.

ZK-STARKs are transparent zero-knowledge proof systems that use collision-resistant hash functions as a security assumption. Their quantum-resistant property applies to the proving layer. Wallet keys and custody arrangements require separate migration work, and Ben-Sasson did not explain how STARKs would fit into wallets.

The National Institute of Standards and Technology (NIST) completed three post-quantum cryptography standards in 2024. FIPS 203 covers general encryption, FIPS 204 covers digital signatures and FIPS 205 covers hash-based digital signatures. Those standards address encryption and signatures, while ZK-STARKs are proof systems.

Ben-Sasson is CEO and co-founder of StarkWare, which develops STARK provers. His post named no AI system that had broken a cryptographic algorithm, no vulnerable wallet and no affected network.

The warning adds a cryptography-focused concern to earlier discussion among crypto leaders about AI-related security risks, including Vitalik Buterin’s warnings about security and privacy risks for crypto users.