Good and bad are morale things. An AI, be it super o hyper or gigapowered, will won't have a morale sense other than what humans put in it (see restrictions about "bad" content in GPT answers). So, an AI is not good nor bad. Eventually if you puts many AI together and allow it to create laws based on what AIs can or can't do, it could lead to artificial morale. But then, nothing tell us if it will be only good morale in a human way (if good morale ever exist).
Today, rational thinking would lead us to abandon oil based technologies, AI included, tbh.
Today, rational thinking would lead us to abandon oil based technologies, AI included, tbh.