At the opening of Devcon 8 in Mumbai, Ethereum co‑founder Vitalik Buterin used a smiley‑mask costume to draw attention to a growing security concern: artificial intelligence could potentially discover mathematical shortcuts that undermine the cryptographic foundations of both Ethereum and Bitcoin, according to a post on X.

ETH / USD
Source: CoinGecko. Prices are indicative.

AI as a new cryptographic threat

Buterin warned that an AI system capable of processing decades of mathematical research in a short time might uncover a shortcut to solve the elliptic curve discrete logarithm problem (ECDLP), the algorithm that secures private keys on both networks. He cited the quadratic sieve—an algorithm from the 1980s that accelerated factoring and weakened RSA—as an example of how a breakthrough can rapidly change the security landscape. While no such shortcut exists for ECDLP today, the possibility of AI‑driven discovery adds urgency to the discussion, BeInCrypto reported.

Beyond quantum computing

The crypto community has long prepared for quantum attacks, planning post‑quantum cryptographic upgrades. Buterin’s comments broaden the threat model to include lattice‑based methods and other post‑quantum standards that could also be compromised by AI. He suggested that algorithms may already exist in unpublished research, and AI could be the tool that brings them to light before developers have a chance to implement mitigations.

Trust and hardware at Devcon 8

In his keynote, Buterin linked the AI warning to a broader agenda for trustworthy infrastructure. He outlined six goals for Ethereum, ranging from formally verified operating systems to self‑sovereign health tech and reliable AI. The emphasis on “trustworthy hardware” echoed similar calls from other speakers, reinforcing the notion that secure hardware, operating systems and AI models are essential to protecting the network’s cryptography.

What to watch next

The Ethereum Foundation is expected to publish updates on post‑quantum and AI‑resistant cryptographic research in the coming months. Stakeholders will be watching for any formal proposals to replace ECDLP‑based signatures, as well as collaborations with academic groups exploring AI‑driven cryptanalysis. Additionally, the broader blockchain community may see renewed investment in hardware security modules and formal verification tools aimed at mitigating both quantum and AI threats.