Discover the Latest Generative AI Tool for Creating Incredible Fantasy Character Art

Literary Lifestyle

Apr 14

Generative AI art has been taking over our feeds for months now. And there’s been a lot of controversy around these images. Some organizations have been accused of training their models on art without the artist's consent, and some people feel it’s outright stealing. Others believe the adage, ‘nothing is original’ and this is just the next wave of art exploration. Either way, you can’t deny a world with ever-developing AI is here to stay. And in the book world, there’s always a desire to see if we can visualize what’s in our heads as we read. I won’t be getting into the companies that are potential bad actors here. Instead let’s see the good, shall we?

Adobe seems to be leading the way in responsible and ethical AI art creation. I finally got access to the beta of Adobe Firefly, and let me just tell you, I’m pretty impressed so far. I had so much fun. Something to note, they make it quite clear they’re working with approved art to train their models:

“We train Firefly using Adobe Stock and other diverse image datasets which have been carefully curated to mitigate against harmful or biased content while also respecting artist’s ownership and intellectual property rights.”

They also mention feedback loops are in place to report any biases so they can continue to improve. In addition, any output from Adobe Firefly is currently not allowed to be used for commercial purposes, and each downloaded image comes with a tag, making it clear it was generated by AI. With all this being said, I felt comfortable with these disclosures, and decided to have some fun. And wow, did I ever.

Inside the beta, there are two options for creation: text to image, and text effects, which turns your letters into the art itself.

To give you an idea of what this generative AI tool is capable of, I've created a few (ok, more than a few!) pieces of art from some of my favorite fantasy series by Sarah J. Maas, as well as others, including Brandon Sanderson’s Mistborn. I used character names with ‘text effects’ and was honestly blown away at the results. You can’t yet input your own images for it to refer to, so all of the images generated are purely text prompt-based. While it definitely struggles with faces, the vibes are immaculate. I’m not going to lie, this is addictive. There are also some unique elements you can control, including depth of field, overhead views, photo or graphic, etc. Below you can swipe through the galleries of the art I generated, and let me know what you think. Have a character you’d like to see? Drop it in the comments.