A few years ago I became seriously ill. I was in a coma on heavy duty meds, and had a kidney transplant. I’m much better than I was, but I can’t do a lot of things like I could before.

We’ve now got quite a few kids in the extended family, so a while ago I wrote a short story to try to make it easier for them to understand. My wife and family like the story and have suggested making it into a picture story book. Problem is, I can’t draw and my imagination isn’t very good.

How can I get pictures for the story if I can’t do it myself and don’t have the money to hire someone? I want to avoid using AI tools because of the potential copyright issues.

I haven’t tried the services like Fiverr because I’ve heard that they force a race to the bottom on prices, but does anyone have any experience, or have any ideas of what I can do please?

Thanks in advance :)

  • rufus@discuss.tchncs.de
    2·
    11 months ago

    Fair enough. I’m pretty sure most of the fuss around AI is way over exaggerated. But we’d need a few more legal disputes and a few new laws to ultimately settle this.

    I’m alright with using AI tools. I think it’s wrong that these companies just take everything they can get hold of, without licensing it. (I mean, I would get in trouble if I downloaded illegal torrents of every novel out there without paying the authors. But it seems companies that develop ChatGPT get away with similar things.) But in the end it’s like if a teacher copies textbooks or shows a pirated version of a movie. This doesn’t make the things the students learned from it ‘illegal’ or forbidden knowledge. I think the same applies to AI. (Given the fact they don’t copy this 1:1 which they usually don’t do anyways… It is possible AI regurgitates its training data in some special cases. So there is some substance to this worry. But I’ve mainly seen this issue come up for example when generating computer code. Less so with images.) But that’s just my oppinion.

    It could bite into your side of copyright, however. But then you’re also embedding that into a context, adding your own text and story around it. Even if the single images turn out not to be copyright-able, the combined work definitely is.

    I don’t want to talk you into using AI. Just: before you end up not doing it at all, maybe re-consider using it. Or do it as a preliminary step to draft something you like. You can still send this version to an artist afterwards.

    And I bet all those issues will be solved in a few years time. Everyone and their grandma are already using AI and AI is not going away. There is no way around that.

    This being said, I think AI also has downsides. Sometimes I generate images or text. But I don’t think it’s proper art. It copies styles well and does what it’s told to do. Sometimes it is super creative and does hilarious things, sometimes the output is a bit bland. But it can’t choose a style for a reason, or embed things into a context or hide little things or a second level of meaning into it’s output. It doesn’t choose colors or other details to underline a certain thing as a proper human artist would do.

    And using AI for real-world purposes is hard. Like you have to learn how to force it to draw the same character in the next image and not a completely different woman. Add the details that fit. Do a consistent stlye. And as I understand people work on their prompts for quite some time before they get the exact results they want. And still, they then generate hundreds of images and go through them to end up with a single good one. It’s way harder to put it to good use, than drawing some astronaut on a horse for some quick fun.