We are thrilled to announce that our campaign to gather artist opt outs has resulted in 78 million artworks being opted out of AI training. This establishes a significant precedent towards realizing our vision of consenting AI, and we are just getting started!

Opt out requests came through individual artists registered through haveibeentrained.com, and also through collaborations with platforms committed to representing the interests of their artists.

@spawning_ As you know, I & many artists are worried that baby steps towards implementing opt-out will serve as political cover for the continued unauthorized use of works to train AI. But congrats here. Let’s fight to ensure this is a step towards accountability, & not the destination.

@spawning_ is there a go-to for understanding how you relate to verifying that material submitted for opt-out or opt-in is created by people submitting it? ps I searched on twitter just now looking for an answer.. and saw you all are saintly amidst misunderstandings in your replies. kudos

@spawning_ Will you publish the list for finetuners wanting to respect it?

@spawning_ The option should have always been opt IN. I'm glad 78 million have pulled but there may well be a lot of work form recently deceased artists who have no way of removing there work. Plus how dose this help with people who have older versions of Stable Diffusion installed?

@spawning_ Alright...so after thinking about this, I'm seeing a huge issue. Opt-in is easy to verify by 3rd parties, because it ALLOWS 3rd parties to access that data. Opt-out does the opposite: the point is to hide/remove urls. How are we going to verify the opt-outs?

@spawning_ Simple question. Why do you not believe in consent? OptOut only worked on people panicking, but legally StabilityAi has no right to the creative works they have gathered. No one should need to be opting out, because none of them had opted in.

@spawning_ In the interest of transparency, can you share numbers of the breakdown here? How many of these were processed by Spawning through haveibeentrained.com, and how many were requests from Shutterstock, ArtStatation, etc?

@spawning_ How do you consider assisting companies in profiteering off of personal and creative data without consent to be the correct route here?

@spawning_ Is your crawler respecting the robot.txt now? It was absolutely not the case before and what about non art related pictures? Also, opt-out it's not the solution at all, Opt-in is the most respectful and fair for everyone.