Artificial Intelligence Tools Could Create Problems and Increase Bugs and Attacks in Cryptocurrency Projects, Warns CertiK Executive

Artificial intelligence (AI) tools such as OpenAI's ChatGPT could potentially introduce more problems, bugs, and attack vectors if used to write smart contracts and build cryptocurrency projects, according to a blockchain security executive. Kang Li, the chief security officer at CertiK, expressed concern that ChatGPT lacks the ability to detect logical code bugs in the same way experienced developers can. This could be particularly detrimental for first-time or amateur-level coders looking to create their own projects. Li suggested that ChatGPT should instead be used as an assistant to engineers, as it excels in explaining the meaning of lines of code. However, he cautioned against relying on the tool for writing code, especially for inexperienced programmers seeking to build monetizable projects. Li expects his assertions to hold true for the next two to three years, although he acknowledged the potential for rapid improvements in AI technology.

AI Tools Becoming More Effective at Social Engineering Attacks

In a related development, Richard Ma, the co-founder and CEO of Web3 security firm Quantstamp, revealed that AI tools are becoming increasingly successful at executing social engineering attacks, often indistinguishable from those performed by humans. Ma noted that his company's clients have reported a growing number of sophisticated social engineering attempts. Some of these attempts involve the use of machine learning to write convincing emails and messages. As a result, it is becoming difficult to discern whether a malicious message is generated by AI or by a human. Ma believes that this trend will continue, with the crypto industry being specifically targeted by AI bots. Hackers can access databases with contact information for key individuals in crypto projects, enabling the AI to send convincing messages in different ways. Ma stressed that it is challenging for companies to train their employees to prevent them from falling victim to these attacks and suggested that better anti-phishing software is needed to mitigate potential threats.