A debate over cryptocurrency’s exposure to advanced artificial intelligence has moved from long-term quantum threats to the possibility that AI could undermine wallet signatures sooner than expected. Justin Drake, an Ethereum researcher, warned that the elliptic curve digital signature algorithm used to secure crypto wallets could fail within months in a worst-case scenario, rather than years. He urged users to move funds gradually to wallets whose public keys have not been exposed.
Haseeb Qureshi, managing partner at Dragonfly, challenged that approach as “cryptographic doomerism.” A wave of theft and forced selling across older addresses could also undermine the value of coins held in newly created wallets, he said.
Qureshi instead proposed a “Cryptographic Recovery Mode” built around a hash-based backup signature that users could associate with their addresses in advance. Under the concept, validators could force a recovery process if the primary cryptographic system failed. The proposal would require blockchain networks to build recovery rules into their protocols before an attack, rather than relying only on users to relocate funds after a vulnerability emerged.
The scale of the exposure is already visible in Bitcoin’s public-key data.
The figures describe visible keys, not stolen funds or an active cryptographic breach. Glassnode’s analysis distinguishes between coins exposed through wallet behavior, such as reusing an address after spending, and coins exposed by the structure of the address itself.
The disagreement leaves blockchain developers weighing three competing responses: gradual user migration, protocol-level recovery mechanisms, and longer-term changes to address designs that reduce public-key exposure.