Co-founder of Ethereum Vitalik Buterin has expressed his excitement about using AI to find possible vulnerabilities in the Ethereum software. Artificial intelligence (AI), according to Ethereum co-founder Vitalik Buterin, may be essential to resolving one of the platform’s “biggest technical” risks: faults hidden deep in its code.
Buterin expressed his excitement for AI-powered audits to find and correct problematic code in the Ethereum network in a post on X on February 18. He called this the “biggest technical risk” to the network.
Buterin’s remarks coincide with Ethereum getting closer to releasing its eagerly anticipated Dencun upgrade, which is presently scheduled for release on March 13. to January 17, Dencun was deployed to the Goerli testnet; however, for four hours, a Prsym problem kept the network from concluding on the testnet. Enhancements to the Ethereum network are essential to the blockchain’s long-term plan.
However, not everyone thinks AI is a trustworthy way to find errors in Ethereum-based software.
In a series of tests carried out in July 2023, OpenZeppelin used OpenAI’s GPT-4 to find security flaws in Solidity smart contracts, which are written in Ethereum’s native language.
In 20 of the 28 tasks that were tested, GPT-4 was able to correctly identify vulnerabilities.
When GPT-4 was unable to find errors, it was frequently asked to make the necessary corrections right away. At other times, though, OpenZeppelin discovered that the AI had created a vulnerability where none had previously been.
In a similar vein, CertiK’s chief security officer, Kang Li, told Cointelegraph that utilizing AI-powered tools in code, like ChatGPT, frequently causes more security problems than it fixes.
Li advises against using AI assistants for any purpose other than helping seasoned programmers, even though they can be useful in helping developers understand what a piece of code implies.
Though Buterin is generally upbeat about AI’s future, he has cautioned developers in the past about using blockchain technology with AI, especially when integrating it with “high-risk” applications like oracles.
It’s crucial to exercise caution because, if someone constructs, say, a prediction market or stablecoin that employs an AI oracle and it turns out that the oracle is vulnerable, a sizable sum of money may vanish overnight.
