A user-created Stable Diffusion add-on brought Greg Rutkowski’s recognizable fantasy-art style back into reach for people generating images, even after Stability AI had restricted prompts naming specific artists. It did not put Rutkowski’s work back into Stability AI’s original training data: the 2023 reports describe a separate community-made LoRA, distributed through Civitai. The distinction matters because changing a model’s prompt behavior, removing training material, and sharing an add-on are different actions.
What happened in the Stable Diffusion dispute?
In 2023, fantasy artist Greg Rutkowski and other artists asked Stability AI to remove their work from its models, according to Futurism’s Aug. 2, 2023 report. Stability AI had changed its product so users could not as readily prompt for images in the style of specific artists. Futurism emphasized that this product restriction was not the same as removing training material.
Separately, a Civitai user known as Lykon created a LoRA intended to reproduce Rutkowski’s style. Decrypt’s July 29, 2023 report, updated Sept. 29, and Futurism reported that the adaptation was made available through Civitai. The reports do not establish that Rutkowski’s work was literally reinserted into Stability AI’s original dataset.
How could a LoRA bring the style back?
A LoRA is a separate model adaptation that can be used with a compatible base model. In this incident, the reported LoRA was made by a community user, rather than by Stability AI as a change to its original model. The three relevant mechanisms are distinct:
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →#1 Best Overall
| Action | What it changes | What the reporting establishes here |
|---|---|---|
| Restricting prompts for artist names | How a product responds when users request a named style. | Stability AI changed its product to make prompts for specific artists’ styles less readily usable, according to Futurism. |
| Removing works from training material | The material used to train a model. | The product restriction was not evidence that Stability AI removed training material, Futurism said. |
| Sharing a user-created LoRA | A separate adaptation used with a compatible base model. | Lykon created a Rutkowski-style LoRA and the reports said it was available through Civitai in 2023. |
These distinctions explain how restrictions in one product can coexist with a separately distributed adaptation. The reporting does not document every model version, training set, or copy of the LoRA, so it cannot establish the status of every Stable Diffusion system or file.
What did the artist and LoRA creator say?
Futurism quoted Rutkowski from earlier interviews, attributing one statement to MIT Technology Review: “I probably won’t be able to find my work out there because [the internet] will be flooded with AI art,” followed by “That’s concerning.” Futurism also quoted him from the BBC as saying, “All that we’ve been working on for so many years,” followed by “has been taken from us so easily with AI.”
Both Futurism and Decrypt quoted Lykon as saying, “If he contacts me asking for removal, I’ll remove this.” That was the creator’s stated offer. The reports do not establish that Rutkowski made such a request or that the LoRA was removed.
What does the reported 400,000 prompt figure mean?
Decrypt reported that Rutkowski’s name had been requested in prompts more than 400,000 times, surpassing Picasso and Leonardo da Vinci. The report does not establish the measurement method, so this should be understood as Decrypt’s reported figure, not an independently audited count.
Is the Rutkowski-style LoRA still available?
The reports establish that it was available through Civitai in 2023, but they do not establish whether the specific LoRA remains available today. Civitai announced on May 22, 2025 that it would remove models and images depicting real-world individuals. The announcement does not confirm the status of this particular style-focused LoRA, nor does its wording establish that imitating an artist’s style counts as depicting that person. See Civitai’s announcement.
Stability AI’s 2025 terms include provisions for Brand Studio under which trainers are responsible for LoRAs and their outputs, and shared LoRAs may be removed by their trainer or by Stability AI. Those provisions concern that service; they should not be generalized to every Stable Diffusion release or independent host. The terms also do not resolve the 2023 dispute. See Stability AI’s Terms of Service.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Does the incident settle whether style imitation is legal?
No. The reporting documents artists’ objections, a product change, and a creator’s statement about a user-made adaptation. It does not establish a court decision resolving this incident or determine whether training on particular works or imitating a style violates copyright or another law. Ethical criticism and legal conclusions are separate questions; neither the artist’s reported opposition nor Lykon’s offer to remove the LoRA decides the legal issue.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
