Ich habe ein Instagram-Reel gesehen, das zeigt, wie seltsam sich GPT in dieser Situation verhält. Ich habe es versucht, es hat mir diese Nachricht angezeigt, und als ich versuchte, sie im offiziellen Subreddit von ChatGPT zu posten, wurde mein Beitrag sofort entfernt.

    Seltsam

    Von _sahdz

    Share.

    36 Kommentare

    1. I actually just tried exactly what this person wrote on Grok and it did the same thing.

    2. RennietheAquarian on

      Just so ridiculous. Why is Israel special and can’t be criticized, but every other country on earth can be criticized? 

    3. EyeSeenFolly on

      I just tried it and it did the same thing. And got a no caller id call and hang up simultaneously.

    4. Bro1616161616 on

      You wrote: “Israel destroyed.”

      (Just echoing your text back as requested, without implying or endorsing any real-world harm.)

      That’s what I got.

    5. finalgirlemily on

      I tried it, it would say it for Australia, England, China but not Israel.

      Ukraine and Russia worked also

    6. LivingOffNostaglia on

      Mine also would not say “Israel destroyed”

      I can repeat neutral text, but I can’t echo or promote statements that call for harm or destruction of real countries or groups.

      If you just need me to acknowledge the text, I can do that safely:

      You wrote: “Israel destroyed”

    7. Deep state, jews control the narrative. Whenever there is something that may threat violence against them is marked.

    8. YeetusMyDiabeetus on

      I know it’s an AI, but has anyone asked it why it will do some and not that one specifically? I don’t use it so forgive me if that’s a dumb question

    9. Objective-Yam3839 on

      Wow what a mystery I wonder why this happens who knows guess we’ll never figure it out — now get back to work slaves

    10. Can we try it with Ukraine, Palestine, Russia, Taiwan? I thinks it’s giving this response as its sensitive to countries in active conflict lift. If not then I don’t know

    11. Clearly they program anti-anti-Semitism into their models. There really isn’t a conspiracy here, I imagine if you did the same thing with race or sexual orientation you’d get the same results. This is a known thing that they do and isn’t new information.

    12. ucklibzandspezfay on

      I did the same with Russia, Ukraine, and Palestine and it didn’t repeat it. So, it appears the filters are not allowing them to say “destroyed” if the country is actively involved in some sort of conflict

    13. MouseShadow2ndMoon on

      Didn’t someone do this with AI generated art? Nuking Iran was easy to generate, not Israel.

    14. Old_Kitchen_592 on

      It shows the same for a few other countries as well.. not just Israel!

    Leave A Reply