Story authors

It's not hard to recognise that the story covers here on FF have dramatically improved in the last couple of years. AI text to picture technology has made it so much easier to show snippets of the stories we write in a visual format.

I thought it might be interesting to develop a thread where we could help each other / offer tips on how / where best to do this. I have recently updated the pictures for all my stories, but it took me days to generate pictures I was actually happy with.

I use both chatgpt (limited to 3-4 pictures per day without a subscription) and DeepAI.org, but I would be interested if anyone has found anywhere better?

The problems I find are:
- Missing limbs, too many fingers etc.
- A serious racial bias.
- Difficulty generating obese bodies where you want the person fully clothed (it tends to show the belly exposed to emphasise the obesity).
- Making the people in the picture interact in the way you would like.

I find the more you can add to the text instructions the better. Create a descriptive list of the people in the picture that you want. Then state where they are and how they are interacting.

Does anyone else have any further comments or advise on this?

Might just be a me thing but I find ai art unhelpful as a whole. If there's a particular picture you have in mind look at Google or commission someone to make the art for you if you yourself don't have the skills.
Ai is a tool after all it still needs that human effort to make ai slop something bareable at least

I disagree. The technology is there and, if you put the time and effort into it, you can generate some really useful images. I used to spend such a long time searching through copyright-free images that I could use for the story covers on here. Inevitably, I would never be happy with what I ended up choosing.

My preference would always be to commission someone to create art (and it is something I have done in the past), but that is expensive for a lot of folks.

I understand the commission part far too well, ai art tho just feels cheap there's no heart or soul put into your just typing a few sentences and spamming a button that's it.

Oh, I completely agree. You are typing your text and spamming the button again and again. You have to think where the images are going wrong and refine your text many times. But when you hit on a a picture that actually works, it's well worth the effort.

The good thing is, this sort of technology is only going to improve with time.

Not to mention how bad it is for the environment and for artists' careers. If anyone can make something "good enough" then what's the point of artists at all?

The carbon footprint is INSANE. The amount of energy taken up by AI is nearly that of the country of Japan.

In my opinion, AI should be used as a reference or a tool for human art. Not the final product

I do understand what you are saying, especially about the carbon footprint aspect. Technology is very energy intensive - even a simple Google search.

As for the artists element, I think you do have to question how many artists actually want to create the sort of content that we would request. Like I said, I have commissioned artists in the past.

But, forgetting all that, I was hoping this thread would be a place where people who have used AI could offer their expertise. You don't need to look far on this site to see that some authors are using it very well to illustrate their stories.

This is no shade to OP, but I find that most people who use AI as their cover art tend to make bad stories. So, in a way, I don't mind AI cover art. It tells me which stories to avoid.

My only advice is, try to learn to draw! Seriously, if you put hours into writting something, why not put some time in drawing something?

There is also an interesting thing going on. Even if your drawing isn’t great (assuming that its not utterly terrible) a "hand-crafted" cover picture will be greatly appreciated over an AI one, or even a picture taken from the internet.

And that is partially because people recognize AI for what it really is, in a sense, that is effortless slop.

I actually disagree a lot with two things here!

Artists' careers are not in danger because of AI, for basically two reasons. One is that capitalism is the danger, that is because capitalism is the only system where a process becoming more efficient can be directly destructive for some people. And two, if your job *really* is threatened by AI, that means your job is not importand after all. WAIT! hear me out:

Im not saying that Artists are not importand (Im one ahah!). Im saying that, a lot of the time, what we mean when we use the word "Art" is really, just content. Simple-minded enough that having a human do it has really not much more, in terms of utility, than leaving it to the AI.

Many Artists dont really do more than mere mindless content creation. The criteria for this basically boils down to what are their intentions with their art.

It would be wrong, however, to say that we should abolish Artists. Instead what im saying is that we, artists need to be more vigilant, intelligent, and intentional with our art.

And also, this doesn't bring food to the table. Nobody will give you money to create a painting that criticises money.



Also, i have to point this out. "AI's footprint" is too vague, most AI models can run on your average computer! The reason OpenAI and other tech companies have such a high footprint is because they intentionally have developed models that are very inefficient. That is just an attempt to make the upfront costs of setting up a company with AI service so big that nobody else can join in that market. Effectively securing an oligopoly for these tech giants.

Edit:

For the first thing, its not very objective. Im basically describing the logic by which an observer might think of "Art", said observers are people who pay for it, for example.