How does AI image synthesis work?

Please see this video.

What is your goal?

I want to show the furry community what can be accomplished with AI and how it can help bring anthro characters to life in new ways.

What AI do you use?

I use Stable Diffusion using Automatic1111 running on my desktop machine.

What GPU specs do you need?

I've seen people run Stable Diffusion with Automatic1111 on a GTX 970 and GTX 1060.

What prompts do you use?

I don't rely on a single prompt or prompt engineering. Most of the images I create rely heavily on compositing and overpainting in Photoshop and inpainting in Automatic1111.

What models do you use?

I'm currently using a mix between YiffAI, Nitro Diffusion, Dreamlike Diffusion, and Stable Diffusion 2.1.

What artists work in Stable Diffusion 2.0?

You can find a list here.

Why SFW? Do you have a NSFW account?

I simply don't enjoy working on NSFW images.

Is there a Discord?

There's the' Furry Diffusion Discord (unrelated to myself).

What is your stance on AI art being theft?

The third pillar of fair use, substantiality, is a key factor in determining whether a work is fair use, or in other words, whether permission from the copyright holder is required. The substantiality of AI art is effectively zero since all output is in some way influenced by each and every image in the training data.