Vitalik Buterin warns AI mathematics could weaken lattice cryptography within two years

"There is a good chance that the concrete security of lattices will take serious hits from the next two years of AI math," Buterin said in a post on X on October 8 ^[15][1]. He said the industry needs to take AI-accelerated attacks on cryptography more seriously, alongside the older threat from quantum computing ^[3]. The cryptotimes.io report is explicit that he is not claiming these systems have already been broken. His concern is that AI could help researchers find mathematical shortcuts that have stayed hidden from humans ^[16].

His case rests on precedent. Over several decades, researchers developed more and more efficient methods for attacking factoring problems, and he said similar discoveries could be waiting for elliptic curves and lattices ^[7]. The factoring record shows that attacks on a structured problem can keep improving for a long time. The thing it doesn't tell you is how large the hits to lattices would be, or when they would come. The report does not include an estimate of lost security for any scheme, or an example of an attack found by an AI system.

Ethereum has been acting on the worry for a year. "This is a major part of the reason why for the past year ethereum's lean roadmap has been going in the 'hash-only' direction," he wrote ^[8]. The approach uses hash-based constructions wherever they can do the required job, so the system depends less on mathematical structures that could contain weaknesses nobody has found yet ^[10]. It has a hard limit. Hashes cannot replace every form of cryptography, and public-key encryption, used in secure communications and privacy systems, needs some mathematical structure ^[11].

I think the evidence supports treating a lattice deployment as provisional, chosen with a plan to replace it if the cryptanalysis moves. It does not show that any lattice scheme in use today needs replacing. Buterin's own forecast is phrased as a chance ^[15].

His wallet advice also favours preparation over haste. "I don't recommend anyone scramble to move their funds to new wallets today," Buterin wrote ^[2]. He added: "If it's not difficult for you, keeping your funds in addresses which have not yet been used to make a transaction is a good idea" ^[13]. Drake, whose proposal came before the post, had suggested that large holders consider moving funds to fresh addresses, and he warned that the process should be controlled ^[14][4].