Using ChatGPT for a "neutral and expert question" is really broken: it gives the public the opinion that ChatGPT output could be trusted whereas it's so often w

Using ChatGPT for a "neutral and expert question" is really broken: it gives the public the opinion that ChatGPT output could be trusted whereas it's so often wrong... That's really dangerous (beyond the immediate stupidity of asking such question to a LLM) because it may give an impression that ChatGPT answers should be trusted and neutral... It's neither.