I see no problem. Everyone assumes Skynet is a murderbot because any pragmatism we can imagine for the robots would demand we be removed. What if there's another pragmatic solution, i.e. the xkcd punchline? Just . . . rewire people to be nice.

Even the robots in The Matrix claim to have tried to give us a utopia, but it made our minds melt down.

I saw that but that was a /pol/ chatbot just regurgitating things people said you don't exactly have to write intelligent comments on /pol/, 90% of the people there write like bots. My guess performance in a slower board like /lit/ or somewhere that requires more knowledge of subject matters and where people write more intelligently it would be called out.

I was messing around with that dudes chatbot OpenAssistant and while interesting it fails the same way all the others do once pressed hard enough (which isn't that hard). These bots seem to do best when you just ask them to regurgitate something from their training library. Which is why non programmers are amazed by it's programming skill (copy pasting from github repo's and stackoverflow) vs devs in the field asking it novel questions and not getting anything interesting out of it.