Made with Bing AI image creator ‘celtic horned god, riding on a horse, at the head of the wild hunt, the landscape is wild and misty, photo realistic, ar 16:9’

I did another attempt this time adding the word ‘Cernunnos’ but I got the annoying and over zealous ‘blocked content’ message…

I mean, I would actually understand it if I was asking it to do ‘tits on a fish’ or something…

  • Thelsim@sh.itjust.worksMEnglish
    3·
    8 months ago

    It looks very majestic!

    Did you ever ask why it blocked the content, or isn’t that something you can do with Bing?
    I had something similar with DALL-E and it’s usually quite forthcoming about why it wasn’t able to create the image. Not that I agree with its reasoning (Swords, and other weapons are OK, but not a laser pistol?) but at least now I know how to circumvent the censorship.

    • quinacridoneOPEnglish
      2·
      8 months ago

      I’ve never asked it, I’m not sure if that is part of the bing function, it does have a button to click if you think that it’s gotten the ‘blocked content’ wrong which I frequently do.

      I’m going to see if I can actually have that conversation with it (I’m still trying to get my head round having a conversation with an ai image creator)

      The weird thing is, when I was making the Aztec god of heart transplants it kept creating a whole range of gory images- the whole chest cavity open, organs falling out etc, which I didn’t want (when I asked it to make sure the chest was closed it got pissy and did the ‘blocked content’ again). It also made a rather nubile Aztec goddess which I wasn’t after either.

      It seems that I can get it to create gore, and t&a by accident, but when I ask for it to calm down a notch, or to provide a seemingly innocuous detail it takes offense!

      • Thelsim@sh.itjust.worksMEnglish
        2·
        8 months ago

        Hmm, I think it’s more the prompting it has a problem with. I’m no expert but in my mind there are two separate AI’s behind the scenes. On the one hand you have Chat-GPT who takes your request and writes a gigantic technical prompt. On the other hand you have DALL-E who takes that technical prompt and tries to create an image based on it. It’s the first one who accidentally puts stuff in the technical prompt and the second one who has a problem with it. With Chat-GPT I just ask it to work around the prompt, suggesting synonyms or other words that might give the same result.
        It’s a bit of a back and forth, but I like having conversations with it :)

        I’m pretty sure there’s no real evaluation of the final image though, only the technical input is being scanned for words on the censor list.

        • quinacridoneOPEnglish
          2·
          8 months ago

          I think I may have a go with the Bing chat in-conjunction with the image creator, just to see what happens…

          I only started using the ai image creator recently, I’m finding it quite fascinating in what it can do, I also showed it to the bf and he was amazed also. He wants to make some original cards for people (Xmas and birthdays)