In the early days of chatGPT you could ask it to list banking websites that are vulnerable to SQL attacks and it would do it. You could ask it to find banking websites with known vulnerabilities and then to write exploits which were usable against the websites.. and it would do it.
4chan do you have any idea how powerful that is? What if we fed it philosophy and asked it to come up with new theories. What if we fed it morality?