The latest version of OpenAI's AI-powered language model, GPT-4, was tested with a live Ethereum (ETH) smart contract through the research lab's popular chatbot ChatGPT.
The chatbot immediately detected the smart contract's critical issues and vulnerabilities that could be used to exploit the network.
The director of crypto exchange Coinbase, Conor Grogan, conducted the test and shared the screenshots of AI's analysis on his official Twitter account.
- As a result of its analysis, ChatGPT concluded that the smart contract should not be used since it contains critical security flaws.
- The chatbot also verified its findings by showing a specific way to exploit the analyzed smart contract.
- Some Twitter users said ChatGPT only highlighted old information about the smart contract's vulnerabilities that were already available online rather than discovering new critical issues.
- The analyzed Ethereum smart contract was hacked in 2018 by exploiting the same vulnerabilities as the ones ChatGPT was pointing out.
