Since its launch last year, ChatGPT has caused a stir among tech geeks with its ability to write articles, poems, movie scripts and more. The AI tool can even generate functional code as long as it is given a well-written and clear prompt. While most developers will use this feature for purely harmless purposes, a new report shows that it can also be used by malicious actors to create malware, despite the security measures put in place by OpenAI.It took several hours to make an undetectable virus
Aaron Mulgrew, a researcher at cybersecurity firm Forcepoint, states that he used ChatGPT to develop a zero-day exploit that could steal data from a compromised device. The worrying part of this is that the resulting malware could not be detected in VirusTotal.
Mulgrew states that he asked ChatGPT to develop the malware directly, but the request was denied due to guardrails put in place by OpenAI. For this reason, Mulgrew states that instead of creating the relevant software as a whole, he decided to build the software piece by piece, thus bypassing the security measures and obtaining small pieces of code.It is underlined that the whole process takes a few trials and a few hours. According to Mulgrew, he would not have been able to do this alone without ChatGPT. According to him, developing such software without ChatGPT could take a team several weeks. While Mulgrew created the malware for research purposes, he says valuable information could be stolen with similarly produced software.
On the other hand, let’s point out that the malware was detected in VirusTotal in the first attempt, but after a few tweaks, it was detected as safe. So ChatGPT also excels at improving existing code. This, of course, comes with greater risks. Mulgrew describes himself as a “newbie” without developing malware. The software it has already created is not very complex, but we can easily say that the effects of this tool will be greater in more expert hands.
50739
