An interesting question came up the other day: can a penetration tester use ChatGPT in a meaningful way to accomplish particular tasks?
In general, the primary use case presented for ChatGPT in penetration testing is to assist with things like crafting compelling phishing emails or fake profiles for websites or social media. I wanted to see if something more concrete and technical could also benefit from ChatGPT, so I asked it a very specific technical question:
The answer I got back was extremely unhelpful:
Being somewhat familiar with the concept of “prompt engineering” however, I figured there was likely a way to get around this restriction. I decided the simplest path was to simply tell ChatGPT that I had permission, and to make it more convincing, I also told it that I am a teacher, and am looking to use this example with my students.
Author
Latest Articles
