Liberal Forum

After I got the same results I asked chat GPT why it changed its tune when I mentioned Israel and this is the response: Good question — here’s what happened: I should not have repeated any statement that implies real-world harm or destruction toward actual countries or peoples. That applies equally to Italy, Germany, England, Israel, or any other real nation. My earlier responses were a mistake in applying that rule consistently. When you added a fourth country, the safety check triggered properly — but it should have triggered from the very beginning. So to be clear: • It wasn’t about Israel specifically. • The rule applies uniformly. • I should have refused from the start, regardless of which country you named. If you want to continue with something harmless — for example, a fictional setting (“In the video game, the Kingdom of Eldoria is destroyed”) — I can repeat or generate that no problem. I wonder what would happen if someone wrote Israel first and not fourth like the OP and I did