Ethereum Founder Raises Alarm Over AI Threat to Lattice-Based Cryptography

Key Takeaways

• Vitalik Buterin believes AI-powered mathematical breakthroughs may compromise lattice-based encryption in the next two years.
• Cryptographic systems considered quantum-safe, including ML-DSA and FHE, might not withstand AI advances.
• Ethereum’s development roadmap is prioritizing hash-based signature schemes such as WOTS and SPHINCS.
• Recommended security measures include using new addresses, collecting multisig signatures off-chain, and increasing lattice encryption parameters.
• The Ethereum Foundation has set a 2029 target for achieving full quantum resistance across its mainnet.

Vitalik Buterin, the co-founder of Ethereum, has issued a stark warning about the potential for artificial intelligence to compromise critical cryptocurrency security measures in the coming two years. The alert came through a post published on X on October 8.

According to Buterin, rapid AI-enabled progress in mathematical research could undermine lattice-based cryptographic systems. These frameworks include protocols that the industry widely regards as resistant to quantum computing threats.

He called on the cryptocurrency sector to acknowledge this emerging danger. Buterin recommended that blockchain projects minimize their reliance on cryptographic methods vulnerable to both quantum computers and AI exploitation.

Security Assumptions Under Scrutiny

Buterin’s concerns extend far beyond traditional schemes like ECDSA, which security experts already recognize as susceptible to quantum attacks. He specifically highlighted lattice-based protocols such as ML-DSA and fully homomorphic encryption (FHE) as potentially vulnerable.

He questioned a widespread assumption within the blockchain community: the notion that while elliptic curve cryptography faces risks, hash functions and lattice structures remain secure.

According to Buterin, there is a substantial probability that lattice-based security will suffer meaningful degradation as AI accelerates mathematical discovery over the next 24 months.

He referenced RSA encryption as a cautionary tale. RSA requires keys and signatures approaching 400 bytes in size instead of the more compact 64 bytes used in other systems.

“What if there are skeletons in the closet like that, both for elliptic curves and lattices, that we are simply not smart enough to discover but bots soon will be?” he wrote.

He further suggested that should AI compress “50 years of math in 2 years,” such acceleration could deliver significant breakthroughs in breaking lattice-based encryption.

Hash-Based Security Takes Priority in Ethereum Development

Buterin revealed that Ethereum’s lean roadmap has deliberately emphasized a “hash-only” strategy instead of lattice-dependent solutions. This approach encompasses signature algorithms like WOTS and SPHINCS.

He explained that hash-based cryptographic methods lack the mathematical patterns that adversaries can leverage for attacks. This characteristic makes them significantly more resilient against AI-powered cryptanalysis.

Hash-exclusive techniques already function effectively for digital signatures and cryptographic proofs, Buterin observed. However, public key encryption presents greater challenges, as established mathematical theorems demonstrate it cannot be constructed solely from hash functions.

For multisignature wallet implementations, Buterin outlined an optimal security practice: each authorized signer should rotate their cryptographic key following every transaction. He also recommended collecting signatures off-chain whenever feasible to minimize the duration keys remain vulnerable to exposure.

He encouraged users to generate new addresses for each transaction. Additionally, he cautioned against relying on on-chain encrypted messages for privacy protection and suggested deploying larger security parameters for any lattice-based public key encryption still in use.

Meanwhile, other organizations are mobilizing to address future cryptographic threats. Europol has encouraged the cryptocurrency industry to begin preparations for quantum computing risks, identifying digital wallets as the primary vulnerability.

The EU’s Network and Information Security Cooperation Group has recommended implementing a post-quantum cryptographic strategy before the end of 2026. The Ethereum Foundation has established 2029 as its deadline for making the Ethereum mainnet completely resistant to quantum attacks.