"If amoral profit maximizers design AI, it's very possibly gonna be an amoral profit maximizer."
The argument from those concerned about AI is richer than that. Any AI, and indeed, any Intelligence in general can be viewed as maximizing its results along a certain gradiant, no matter how complicated that gradient may be. If an AI determines that improving itself would help it achieve its goals, and there isn't some sort of natural limit to intelligence that would stop it, the first one to develop will improve itself until it is exceedingly good at maximizing that metric, whatever it may be. Even if someone quite altruistic comes along and programs this first AI to "maximize the happiness of humans", and gets even a slight bit sloppy with their gradient, they may create an AI that optimizes the universe for human happiness by converting the entire universe into human pleasure centers being stimulated at the maximum possible rate, and nothing else that you might be concerned about, such as your forebrain or visual center or anything else that you may think makes you human, but the highly altruistic and very well-meaning AI programmer neglected to specify in their definition of human.
Here's another way of looking at the AI concerns: Let us stipulate a magical force in the universe that constrains AIs to be "fairly benevolent". Can you, right now, mathematically specify what a human is, and mathematically specify the best thing for humanity, in such a way that you are willing to 100% commit yourself and all humans to those definitions, rigidly, and for all time? And are you comfortable with the fact that it probably won't be you deciding?
For context, I'm not necessarily a big believer in this argument, but that doesn't mean it isn't worth at least thinking about, and characterizing correctly.