A tool is just that a tool. Any tool can be used for good or for evil purposes. It's hard to imagine that a well regulated AI is all that is needed to get the Chinese government to stop brutally oppressing the Uygher people. Regulate AI all you want, it won't stop nasty people from doing nasty things. It will stop bitemenow999 from making his cat dog model. It will stop a lot of very productive people from doing important and positive work with AI.

If a graduate student no longer wants to peruse ML because of his own moral code, that is his choice. There is no reason that I, or anyone else, should be regulated from doing research in this area because of someone else's hang ups.