I have no real ability to create text prompts that match an image in my head. It’s usually faster for me to use my own photos or some I find on Unsplash and remix them in Photoshop than it is to get an AI result. So, I thought it might be fun to compare free generators.
Prompt: Ancient Galilee young Jewish husband and wife facing each other in thoughtful conversation.
I had to create accounts for all of them, but could generate a couple of iterations for each before running out of credits/tokens. I think Fotor and Imagine did the best. Some of the generators ignored “ancient” and others ignored “Jewish.” (I doubt very many of my modern Jewish friends have crucifixes on their walls, much less any first-century people of any faith tradition!) Hands and feet are hard to draw, even for AI. Check out the feet of the poor man in the NightCafe version. I’m not sure what references Firefly/Adobe uses, but this image explains why I have so little success with “generative fill” in Photoshop. This couple is not young and they are not looking at each other. At least the woman is attractive, unlike the woman with a bulbous nose that Canva created.
After playing with a realism(ish) goal, I thought it might be fun to test out some fantasy.
Prompt: Woman smiling in a colorful canyon with wolves ready to attack behind her
Fotor was the only one to get the prompts correct, but I love what NightCafe did (@sereid, do you you see those??). As before, the Firefly/Adobe version is awful. Pity, since I already pay for Adobe. Fotor is too expensive at $40/year, but NightCafe is pretty reasonable and you can earn credits by interacting with the community. There is no mistaking AI for real art, but it does help me generate ideas for my own digital artwork and stories, too.
What are your thoughts on AI illustration?