I don't think that's modeling the risk correctly. In my mind the risk is that ChatGPT's creators are able to influence your political opinions _without_ you seeking that out