# Vitalik Buterin Advocates for AI-Powered Verification to Make Crypto Safer

*Author: Jack Inabinet*
*Published: May 18, 2026*
*Source: https://www.bankless.com/de/read/news/https-x-com-vitalikbuterin-status-2056354141832626487-s-20-advocates-for-ai-powered-verification-to-make-crypto-safer*

---

While much of the conversation around artificial intelligence and cybersecurity has centered on how artificial intelligence makes vulnerabilities easier to discover, Ethereum co-founder Vitalik Buterin is taking more contrarian view.

### What's the Scoop?
- **A/ACC Cybersecurity: **In a newly published [blog post](https://vitalik.eth.limo/general/2026/05/18/fv.html), Vitalik Buterin argues that AI-assisted “formal verification” will become one of cybersecurity’s most important tools, accelerating the adoption of machine-verifiable mathematical proofs to ensure software behaves exactly as intended. While Buterin says that increasingly powerful AI systems (like Claude Mythos) can dramatically accelerate vulnerability exploits, he posits that the combination of formal verification and AI will supercharge vulnerability discovery before exploits ever occur, helping developers ship secure open source code.
- **Crucial Caveats:** Buterin stresses that formal verification is not a silver bullet. Proof systems can miss critical assumptions, fail to capture real-world attack surfaces, or overlook vulnerabilities in hardware and adjacent systems. Additionally, AI can lower the barrier to generating large amounts of software, increasing the risk that insecure or poorly understood code enters production systems. Still, Buterin argues the AI-assisted verification meaningfully reduces catastrophic risk compared to traditional software development processes.
- **Provable Infrastructure:** Multiple projects are already applying formal methods to Ethereum’s stack. Buterin highlights efforts like Arklib, which aims to build a fully formally verified STARK implementation, alongside formally verified EVM implementations written directly in RISC-V assembly.

> Many people have claimed that with AI-assisted bug finding, secure code (and hence trustless anything) will be impossible.I have a much more optimistic take, and AI-assisted formal verification is a major part of the reason why:[https://t.co/0ceMBZ6uqj](https://t.co/0ceMBZ6uqj)— vitalik.eth (@VitalikButerin) [May 18, 2026](https://twitter.com/VitalikButerin/status/2056354141832626487?ref_src=twsrc%5Etfw)