After playing around for a bit with the current tools for AI image generation, I’ve found that AI is more competent than I thought at generating graphics for things like posters or social media posts. What really surprised me was how much better text generation has gotten. I tested a Chinese image generation tool called “Dreamina” and also Dall-E. Dreamina is genuinely miles ahead of Dall-E (at least the version I used) to the point where if you weren’t looking too closely, you might not be able to tell that the stuff produced by Dreamina is AI generated. Dall-E’s results had such a blatant AI look.

But what do I really think about this? You know the genre of (I’m not trying to gatekeep!) “casual” graphic designers; people who use software like Canva to make stuff for their job or for school/university and who aren’t exactly well versed with the theory side of graphic design (that’s not a bad thing!)? I think AI tools are useful for that kind of designer. The results I got from Dreamina for things like social media posts were surprisingly good, but they looked wholly like the Canva kind of social media post.

However, I don’t think AI tools, at least not soon, are replacing professional or “artist” graphic designers. For example, I have zero faith in AI tools being able to execute a consistent visual identity for a business, at least right now. That, and also people who graphic design for the “art” of it (the whole argument about AI art being “soul-less”) aren’t going anywhere.

In my opinion, I don’t believe AI should be used to do all of the work in this regard (as with other stuff too), but it could be used as a tool to do some of the more boring stuff (though I haven’t started doing that yet). Maybe, stuff like generating backgrounds, or finding insipiration, or quickly testing ideas you may have without having to commit a lot of time. In that sort of sense, I can see a use for AI.

Something I found with Dreamina, and I’d imagine similar tools, is that if you don’t like something about the result its given you, you kind of have to fight against it and hope for the best. Dreamina has a feature where you can tell it to remove something from an image, or change something about it. Both might work, but they don’t seem to be consistent. Sometimes the feature where you can change stuff about an image just added a bunch of classic AI deformed text into the image instead of what I told it to do!

That leads me to another point: as someone who can use graphic design software, and as someone who’ll 99% want to change some stuff about whatever these AI tools might produce, I can’t change stuff myself! These tools produce their output as an image, and I can’t change text in an image! For me, this renders these tools practically useless! I could imagine using AI tools for graphic design work I don’t want to do (the kind of work you do to just get done with it), but if you can’t edit anything without having to leave it to luck, then there’s no point for me! Something that can generate a file that you can edit yourself in other software might prove more useful.