Vitalik: Ethereum to Become Key Part of 'Security Core,' AI-Assisted Formal Verification Can Enhance Code Security

May 19: Ethereum co-founder Vitalik Buterin dropped a deep dive on the current state and future of formal verification tech. He argues that pairing formal verification with AI will emerge as the “ultimate form of software development” — and a critical way to protect network security from powerful AI-driven threats. Vitalik explains that formal verification’s core boils down to turning code correctness into a mathematical theorem that computers can check automatically. This ensures software security via mathematical proofs, not just traditional testing or manual audits. He says this approach works especially well for cases where the “goal is far simpler than how you build it” — like quantum-resistant signatures, STARK proof systems, Byzantine fault-tolerant consensus algorithms, and ZK-EVM. These are exactly the core tech components for Ethereum’s next phase of upgrades. That said, Vitalik admits formal verification isn’t a cure-all. Common pitfalls include checking only parts of code and missing critical flaws, built-in gaps in security specs, and side-channel attacks at the software-hardware interface that existing proof models can’t easily catch. He stresses that “provable correctness” is all about verifying how well different versions of an intent align — not whether code is perfectly correct against what humans actually want. When it comes to AI’s network security threats, Vitalik is cautiously optimistic. He thinks AI-aided formal verification will push software architecture to evolve into a “secure core, insecure outer layer” model: edge components run in low-permission sandboxes, while the secure core is constantly reinforced via formal methods — taking on the biggest trust burden as society becomes more digital. Ethereum will be a key part of that secure core.