AI Could Break Encryption Algorithms, Ethereum Researchers Call for "Bunker Mode"

On October 14 that, according to Cointelegraph, Ethereum researcher Justin Drake warned that rapid advancements in AI could accelerate mathematical attacks on elliptic curve digital signature algorithms (ECDSA), potentially breaking existing encryption frameworks within months under the worst-case scenario. He urged the industry to activate "bunker mode"—transferring assets to new addresses where the public key has not been exposed—to reduce risks.

Several industry insiders remain skeptical in response. Coinbase Chief Cryptographer Yehuda Lindell noted that there is currently "no evidence" indicating the mathematical foundations of ECDSA have been compromised; Trezor hardware wallet Bitcoin analyst Lucien Bourdon also cautioned that migrating assets could backfire, inadvertently providing openings for fraud. MetaMask founder Taylor Monahan bluntly added that if elliptic curve cryptography were actually breached, "the entire world would go to zero," making bunker mode ineffective.

Ethereum has already developed a post-quantum roadmap based on hash-based signatures and ZK proofs, while public chains such as Solana and NEAR are also exploring their own quantum-safe solutions, though none have announced specific implementation timelines yet.