CanadianKingpin12 is actively training new chatbots using unrestricted data sets from the dark web or large language models developed for fighting cybercrime. The individual also has access to another large language model, DarkBERT, developed by South Korean researchers. Despite being intended for cybercrime prevention, DarkBERT is being used for malicious purposes. The trend of using AI chatbots for cybercrime is growing, with hackers having access to advanced tools for executing sophisticated attacks.
Key takeaways:
- Cybercriminals are training AI chatbots for malicious purposes such as phishing, social engineering, exploiting vulnerabilities, and creating malware. Notable tools include FraudGPT and a new tool under development based on Google's AI experiment, Bard.
- The same individual, known as CanadianKingpin12, is behind both AI-powered bots. FraudGPT was released on July 25 and is being promoted on various hacker forums.
- CanadianKingpin12 is also working on DarkBART, a "dark version" of Google's AI chatbot, and has access to another large language model named DarkBERT, developed by South Korean researchers to fight cybercrime but is being used for malicious purposes.
- The trend of using generative AI chatbots for cybercrime is growing, providing an easy solution for less capable threat actors or for those wanting to expand operations to other regions and lack the language skills. This development underscores the significant influence of malicious AI on the cybersecurity and cybercrime landscape.