According to Buterin, the application of AI in formal code review could streamline the process and mitigate significant technical risks, ultimately enhancing the functionality of the Ethereum network.
One application of AI that I am excited about is AI-assisted formal verification of code and bug finding.
Right now ethereum's biggest technical risk is probably bugs in code, and anything that could significantly change the game on that would be amazing.
— vitalik.eth (@VitalikButerin) February 19, 2024
Buterin’s remarks precede the highly anticipated Dencun update scheduled for launch on March 13. While Dencun was successfully deployed on the Goerli testnet on January 17, a bug in the Prysm client caused a four-hour delay in operations.
However, not everyone shares the belief that AI is a reliable tool for bug detection in Ethereum code. OpenZeppelin carried out experiments in July 2023 using OpenAI’s GPT-4 to identify security issues in Solidity smart contracts. Although GPT-4 successfully detected vulnerabilities in 20 out of 28 tasks, OpenZeppelin later discovered that the AI had actually invented a non-existent vulnerability.
CertiK Security Director Kang Li cautioned against relying solely on AI-based tools like ChatGPT for coding security, as they often introduce more issues than they solve. Li recommended using AI as an assistant alongside experienced programmers engaged in code analysis and reverse engineering.
Recently, Buterin also issued a warning to developers considering AI adoption for prediction markets or stablecoins. In November, the Ethereum co-founder expressed concerns that artificial intelligence could turn against humanity, potentially resulting in significant financial losses for individuals.