Yes and no. The ability to "understand" code that makes AI capable of finding security bugs, is correlated with AI's ability to write code. AI is just really good at code related tasks, and categorically banning all AI usage is shooting yourself in the foot.