Malware may use ChatGPT to rewrite its own code
Jusan/Getty
The researchers found that computer viruses can use ChatGPT to rewrite code to evade detection, create customized emails that look like genuine replies, and spread via email attachments.
Large-scale language models (LLMs), the artificial intelligence behind powerful chatbots like ChatGPT, can not only generate human-like text but also write computer code. David Zollicoffer Researchers at ETH Zurich in Switzerland and Benjamin Zimmerman of Ohio State University are concerned that this feature could be exploited by so-called metamorphic malware, viruses that rewrite their own code.
Source: www.newscientist.com
