Artificial intelligence technologies, such as OpenAI’s ChatGPT, would cause more problems, defects, and attack vectors if used to design smart contracts and establish cryptocurrency projects, according to a CertiK executive.
During Korean Blockchain Week on September 5th, Kang Li, the chief security officer of CertiK, shared with Cointelegraph that ChatGPT cannot identify logical code bugs as effectively as experienced developers do.
Li expressed concerns about the potential drawbacks of using ChatGPT. They mentioned that relying on this tool might increase software bugs rather than helping to identify and fix them. This could be particularly problematic for individuals who are new to coding or have limited experience, as it could negatively impact their ability to develop their projects successfully.
ChatGPT opens up opportunities for individuals who need more training to participate. They can begin using it immediately, but I have concerns about potential hidden issues related to morphological design. According to him, when you use ChatGPT to assist you in creating something, there is a risk of it failing badly due to specific design flaws, especially when faced with attacks from malicious individuals.
On the other hand, Li thinks that ChatGPT can be a valuable tool for engineers, specifically as an assistant. According to Li, one of the strengths of ChatGPT lie in its ability to explain the meaning behind a particular line of code effectively. ChatGPT is a handy tool for individuals engaged in code analysis and reverse engineering. This assistant is truly impressive and will significantly enhance our productivity.”
He emphasized the importance of not depending on it for coding, especially for novice programmers who aim to create something profitable.
Li expressed his commitment to support his claims for at least two to three years, recognizing the potential for significant advancements in AI that could significantly enhance ChatGPT’s abilities. During an interview with Cointelegraph at KBW on September 4th, Richard Ma, the co-founder and CEO of Quantstamp, a Web3 security firm, discussed the increasing effectiveness of AI tools in social engineering attacks. Ma pointed out that these AI-driven attacks often resemble the tactics employed by humans.
According to my mother, Quantstamp’s customers have been experiencing a concerning increase in highly advanced social engineering attacks. I’ve observed lately that individuals have employed machine-learning techniques to compose emails and messages. The current attempts at social engineering are much more convincing than those from a few years back.
For quite some time now, regular internet users have been dealing with the nuisance of receiving AI-generated spam emails. However, Ma believes that we are reaching a critical juncture where it will become increasingly difficult to distinguish between malicious messages created by AI and those crafted by humans. “It will become increasingly challenging to differentiate between messages from actual humans and those from highly convincing AI, especially when it comes to personal messages,” he remarked. Experts in the crypto industry are already facing targeted attacks, with some individuals being impersonated by artificial intelligence bots. Ma holds the belief that the situation will deteriorate further.
Numerous cryptocurrency databases contain contact information for each project’s key individuals. Hackers can gain access to that information and utilize an artificial intelligence system that enables them to communicate with people using various methods.
Ma emphasized the difficulty of training an entire company to refrain from responding to such matters. According to my mother, a new, improved anti-phishing software will soon be available. This software is designed to assist companies in reducing the risks of potential phishing attacks.
