Vitalik Buterin, who helped create Ethereum, has addressed growing worries that using artificial intelligence to find flaws in code could flood developers with issues and lead to constant hacking attempts on blockchains.
He believes this technology could soon strengthen the security of cryptocurrencies. He explains that using AI to formally verify code might become a powerful way to prevent security problems in crypto and across the internet.
AI Could Strengthen Security Instead of Breaking It
Formal verification involves using computers to automatically check if software works correctly by creating mathematical proofs, rather than relying on human reviewers. While this technique has existed for a long time, it hasn’t become widely adopted because creating those proofs by hand was time-consuming and difficult for developers.
Vitalik Buterin now believes AI is changing how developers work with blockchain security. Instead of manually creating the code and the mathematical proofs that verify it, developers can now use AI to generate both. They can then simply confirm that the AI has proven what they intended.
As a researcher in this field, I’ve been considering a potentially disruptive scenario. Imagine AI reaching a point where it can automatically identify vulnerabilities in existing software code. This raises a critical question: what are the implications for systems where even a single flaw could have catastrophic consequences for users – potentially costing them everything?
As a crypto investor, I was really interested in his explanation of formal verification. Basically, it’s a way to mathematically *prove* that code does exactly what it’s supposed to. The cool part is, if you use it throughout the entire system, a powerful AI trying to find bugs would already be looking at code that’s been verified – meaning the major flaws would already be ironed out. It’s like a pre-emptive strike against vulnerabilities.
He highlighted a couple of Ethereum projects already exploring this method. Arklib is focused on creating a completely verified STARK implementation. Meanwhile, evm-asm is developing an EVM using low-level RISC-V assembly and ensuring it works correctly by comparing it to a clear, understandable version.
Vitalik Buterin noted that both Claude and Deepseek 4 Pro are capable AI models for writing Lean proofs.
He also pointed out that Leanstral, a smaller AI model designed for the Lean programming language, can run on a standard computer and actually perform better than much larger, more general AI models when tested on tasks that require proving the correctness of code.
But There Are Limitations
As an analyst, I noted that while Vitalik Buterin is a proponent of formal verification, he also spent a significant portion of his writing detailing its practical limitations and instances where it hasn’t quite lived up to expectations.
This covers a range of issues, like errors in even reliable compilers, problems in code libraries where only some parts were checked for errors, and designs that were technically correct but didn’t actually meet the developer’s intended goals.
He sees formal verification not as a complete solution for security, but as a strong addition to the ongoing effort to reduce errors in software. It’s one piece of a larger movement toward creating more reliable code.
It’s important to understand the context: Vitalik Buterin’s post came at a time when the cryptocurrency world was already shaken. Just four days prior, a hacker stole over $76 million worth of cryptocurrency from the Echo Protocol, marking the third major security breach in that short period.
Just days ago, news broke about a cyberattack on THORChain, resulting in losses exceeding $10 million for the platform.
It wasn’t long before another hack hit the crypto space, this time targeting the Verus-Ethereum Bridge. Someone managed to steal $11.58 million because the system didn’t properly verify something it should have. Honestly, as an investor, it’s frustrating because this is exactly the kind of specific, easily avoidable mistake that a more rigorous security check – like a formal proof – could have potentially prevented.
Read More
- Pi Hotel Vietnam: First to Accept Pi Coin Payments in Real-World Transactions
- Silver Rate Forecast
- Gold Rate Forecast
- USD TRY PREDICTION
- Warsh’s Fed Debut: A June Rate Cut? Don’t Hold Your Breath, Darling
- 3 MASSIVE Token Unlocks That Could Blow Up the Market: HYPE, ENA, RED!
- RENDER PREDICTION. RENDER cryptocurrency
- AAVE’s New Yield Revolution: Can It Keep The Fresh Buyers Coming?
- USD IDR PREDICTION
- SOL GBP PREDICTION. SOL cryptocurrency
2026-05-20 00:43