AI’s Potential to Address Ethereum’s Technical Vulnerability

Ethereum’s co-founder, Vitalik Buterin, believes that artificial intelligence (AI) could play a crucial role in addressing one of Ethereum’s major technical challenges — hidden bugs in its code. Buterin expressed his excitement about using AI-powered auditing to identify and fix these bugs, which he considers the most significant technical risk to the Ethereum network. His comments come as Ethereum prepares to launch its highly anticipated Dencun upgrade on March 13. Although Dencun was successfully implemented on the testnet, a bug in Prsym caused a delay of four hours in finalizing the network.

Upgrades to the Ethereum network are essential for the blockchain’s long-term roadmap. Not everyone agrees on the reliability of using AI to detect bugs in Ethereum-based code. OpenZeppelin conducted experiments in July 2023, using OpenAI’s GPT-4 to identify security issues in Solidity smart contracts, the native language of Ethereum code. The results showed that GPT-4 successfully identified vulnerabilities in 20 out of 28 challenges. When it failed to recognize flaws, it could be prompted to correct its mistakes quickly. There were instances where the AI invented vulnerabilities that did not exist in the first place.

Kang Li, the chief security officer at CertiK, also expressed skepticism regarding the use of AI-powered tools like ChatGPT when coding. Li believes that these tools can create more security issues than they solve. Instead, he recommends using AI assistants as a supplementary tool for experienced coders to quickly explain lines of code. He acknowledges that ChatGPT can be a beneficial assistant, improving developers’ efficiency in code analysis and reverse engineering.

While Buterin remains optimistic about the future of AI, he has previously cautioned developers about the potential risks of implementing AI with blockchain technology, especially in high-risk applications like oracles. He emphasized the importance of being cautious, highlighting the potential loss of significant amounts of money if an AI oracle used in applications such as prediction markets or stablecoins were to be vulnerable to attacks.

Ethereum’s co-founder is enthusiastic about the potential of AI to tackle bugs in their code. There are differing opinions on the reliability of AI tools when it comes to identifying vulnerabilities in Ethereum-based code. It is crucial to exercise caution and use AI assistants as supplementary tools for experienced coders, rather than relying solely on AI for code analysis and security.