Implementation for Malicious Software Using ChatGPT-4

Large language models, including ChatGPT, have become widely available and have been used in various fields. This study attempts to create malicious software using ChatGPT-4 with minimal coding effort to demonstrate that they can be created even by non-expert security professionals. We also show whether obfuscating the source code of the created malicious software can reduce the malware detection rates by security analysis tools.