What jimrandomh is talking about is not just whether megacorps use AI for good or bad, but what happens if an AI starts doing damage outside the control of anybody. That could happen regardless of democratization if advancement occurs too quickly without thinking about AI safety.