The Malicious Evolution of ChatGPT: How Computer Viruses are Using AI to Spread Metamorphic Malware through Email Attachments
David Zollikofer from ETH Zurich and Benjamin Zimmerman from Ohio State University are concerned that computer viruses have the ability to utilize ChatGPT to rewrite their code, ultimately spreading themselves through email attachments. Along with generating human-like text, large language models (LLMs) such as those behind ChatGPT can also write computer code. This capability could be leveraged by viruses that modify their code, a type of malware known as metamorphic malware. Researchers have demonstrated that this allows the virus to create customized emails that appear to be genuine replies, making it harder for users to detect and prevent the spread of the virus.