Blockchain security experts debate AI and quantum risks to cryptography

Ethereum researcher Justin Drake has warned that advances in artificial intelligence could accelerate the timeline for breaking the elliptic curve digital signature algorithm (ECDSA), which secures many cryptocurrency wallets. Drake suggested that in a worst-case scenario, the timeframe for such a breach could be measured in months rather than years, citing recent progress in AI-driven mathematics.

In response to these concerns, Dragonfly managing partner Haseeb Qureshi has criticized a narrow “bunker mode” approach that focuses solely on moving funds to new addresses. Qureshi argues that such individual actions are insufficient if the broader network remains vulnerable. He proposed a “Cryptographic Recovery Mode” to allow validators to recover funds through hash-based backup signatures.

Bitcoin Core contributor Antoine Poinsot has also examined these risks, distinguishing between AI-assisted software flaw discovery and the potential for quantum computing to break fundamental cryptography. Discussion among developers includes potential protocol changes, such as BIP 360, BIP 361, and P2TR v2, to mitigate risks like the theft of funds or the freezing of coins. Implementing these safeguards remains subject to community consensus and the practical challenge of migrating funds from exposed addresses.