Ethereum co-founder Vitalik Buterin has indicated that artificial intelligence may compromise specific cryptographic standards within the next two years. In a post on X, Buterin described it as "plausible" that AI-accelerated mathematics could significantly weaken lattice-based cryptography, a method often proposed for quantum-proofing blockchains. This warning follows comments by Ethereum Foundation researcher Justin Drake, who suggested that AI might break encryption before quantum computing reaches the so-called Q-Day threshold.
Buterin emphasized that while he does not recommend scrambling to move funds immediately, the industry must minimize exposure to both quantum-vulnerable and potentially AI-vulnerable cryptography. He noted that projects like Cardano have considered lattice-based solutions, but shifting assets to these addresses could inadvertently expose users to short-term AI risks. Isabel Foxen Duke of the Mara Foundation supported this view, stating that classical breaks via AI could affect cryptography broadly considered quantum secure, necessitating more conservative approaches. Buterin suggested increasing lattice sizes or using hash-based cryptography, though he acknowledged AI could theoretically impact hashes as well.
The convergence of AI capabilities and cryptographic vulnerability introduces a complex layer to blockchain security planning. Traditionally, the primary threat model focused on quantum computing breaking elliptic curve and RSA cryptography. However, Buterin’s assessment suggests that AI-driven mathematical breakthroughs could accelerate the degradation of security margins for lattice-based systems, which are currently viewed as a leading candidate for post-quantum resilience. This creates a strategic dilemma for networks attempting to future-proof their infrastructure: adopting quantum-resistant methods may expose them to immediate, shorter-term risks from advanced computational techniques powered by AI.
Market participants and protocol developers must now navigate a dual-threat landscape where the timeline for cryptographic obsolescence is compressed. The advice to keep funds in addresses with unexposed public keys highlights a practical mitigation strategy that relies on operational hygiene rather than immediate technological overhaul. This stance underscores the importance of cautious, incremental upgrades over reactive panic. As AI models evolve, the assumption that certain cryptographic primitives remain secure for decades may no longer hold, requiring continuous reassessment of security assumptions and potential adjustments to consensus mechanisms and wallet protocols.
