OWASP found that ChatGPT helped an operator with no offensive security knowledge take over an administrator account during a controlled hacking test. The operator largely copied commands, returned results to the chatbot, and followed its guidance to change a password and log in. The finding shows...