On Friday, online art community DeviantArt announced DreamUp, an AI-powered text-to-image generation service powered by Stable Diffusion. Simultaneously, DeviantArt launched an initiative that ostensibly allows artists to opt out of AI image training, but also forced everyone’s art to enroll by default, who irritated many members.
DreamUp creates new AI-generated art based on text prompts. Due to its stable streaming roots, DreamUp learned how to generate images by analyzing hundreds of millions of images pulled from sites like DeviantArt and collected into LAION datasets without permission from the artists, a potential irony that some members by DeviantArt find problematic.
this is how dreamup is developed. it adds non-opt-out illustration training on top of the stability diffusion model, which is ALREADY trained using unethical data. they cannot unlearn these parts.
this whole discussion is not in good faith unless they address this pic.twitter.com/KnQZerCIol
— svlt ✈️ (@svltart) November 11, 2022
As we’ve often reported on Ars in the past, the web-scraping nature of Stable Diffusion sparked a huge debate earlier this year among artists who challenge the ethics of AI-generated artwork. Some art communities have taken strong stances against all AI-generated imagery, banning it altogether.
Perhaps anticipating a backlash, DeviantArt is making overtures to appease artists who might be upset that their work is being used to train AI image generators. The site provides a special “noai” flag that artists can check in their image settings to opt out of third-party image datasets. (However, it remains to be seen whether third-party image scrapers will honor this flag.)
Additionally, DeviantArt will allow artists to not let their images form DreamUp in the future, but each artist must first fill out a form that requires human review. This policy has resulted in significant pushback among DeviantArt members, some of whom have threatens to delete all their work and deactivate their accounts.
DeviantArt’s DreamUp information page also takes a defensive tone, stating that DeviantArt has not consented to third-party AI image models (such as Stable Diffusion) that have scraped their site to run their models. And further down the page, the site attempts to debunk common misconceptions about how AI image synthesis works.
As for DreamUp itself? We experimented with the service, which looks like a Vanilla Stable Diffusion model. DeviantArt members get five free invites to try it out, and members can earn more invite credits by subscribing to various CORE plans ranging from US$3.95 to US$14.95 per month.
Alternatively, you can also use Stable Diffusion locally for free if you’re handy enough to install a package from GitHub or download the Draw Things app on your iPhone.