Vitalik Buterin: AI may weaken cryptography within two years and put ECDSA at risk, suggesting a priority on hash cryptography

Ethereum co-founder Vitalik Buterin posted on the X platform, stating that AI-accelerated mathematical research may pose new security threats to existing cryptographic systems, especially lattice-based cryptographic schemes, including ML-DSA and fully homomorphic encryption (FHE). He believes that advancements in AI in the field of mathematics over the next two years could significantly undermine the practical security of lattice cryptography, and ECDSA may face the risk of being compromised sooner than expected.

Vitalik pointed out that over the past year, the Ethereum Lean roadmap has gradually shifted towards pure hash cryptographic schemes, avoiding the use of lattice-based signatures and related zero-knowledge proof structures. He suggested prioritizing the use of hash cryptography whenever feasible and adopting more conservative security parameters for lattice cryptography; privacy protocols should avoid storing encrypted information directly on-chain, and multi-signature wallets should prioritize off-chain signature confirmations. Additionally, Vitalik advised users, when operationally convenient, to store funds in addresses that have not initiated transactions to reduce potential risks, but he does not recommend users immediately migrate assets on a large scale. He emphasized that the migration process itself also carries risks, and his own losses due to migration errors have even exceeded the losses from past hacking attacks.