AI tools for artists.
Made by artists.
About
Spawning is building tools for managing your AI identity.
We believe that a future of consenting data will benefit both AI development and the people it is trained on.
You may have used our first tool haveibeentrained.com to see if you are present in popular training datasets, and opt out or opt in to future training, or have seen Holly+, the first project to experiment with consensual interactions around an artist AI model.
We have a lot more in the works.
Team
︎ Mathew Dryhurst
︎ Holly Herndon
︎ Patrick Hoepner
︎ Jordan Meyer
︎ Cullen Miller
︎ Nick Padgett
︎︎︎ @matdryhurst
︎︎︎ @hollyherndon
︎︎︎@ExquisiteCorps5
︎︎︎ @jordancmeyer
︎︎︎ @cullenmiller
︎︎︎ @ncpadge
We are a group of artists, scientists and engineers with notable backgrounds in machine learning and art. We have experimented with consenting model training in art projects for many years and formed Spawning to steer the rapid commercial development in large language, image and music models towards consent.
What does Spawning mean?
Spawning is a term we created to describe the act of creating entirely new media with an AI system trained on older media. We felt it was necessary to distinguish that this process is different from older techniques like sampling or collage.
What do you mean by opting in or out of
AI models?
AI models?
AI spawning systems like OpenAI’s DALL·E or Stable Diffusion are trained on billions of images.
These systems allow for anyone to generate artwork in the style of artists and individuals who are well represented in their training set. If you would like to see if your art or likeness is currently in popular AI training datasets, you can use our tool haveibeentrained?, and register to use our opt-in and opt-out tools in beta.
We believe that both the AI research community and the artist community will benefit from this training being consensual.
These systems allow for anyone to generate artwork in the style of artists and individuals who are well represented in their training set. If you would like to see if your art or likeness is currently in popular AI training datasets, you can use our tool haveibeentrained?, and register to use our opt-in and opt-out tools in beta.
We believe that both the AI research community and the artist community will benefit from this training being consensual.
What kind of artists can participate?
All kinds of artists. Whether you are a visual artist, musician, designer, writer or filmmaker, we feel our tools can be beneficial for you.
Do you believe in a future of copyright warnings and DMCA takedowns?
There are no guarantees that copyright will be sufficient to protect artists from AI training, so we have little choice but to operate assuming it will not.
While legal proceedings that determine the fate of copyrighted works are out of our hands, what is within our grasp is the ability to conceive of a new era of consenting interactions around artist data.
A new era offers us the opportunity to reconfigure how we treat IP! We believe that the best path forward is to offer individual artists tools to manage their style and likenesses, and determine their own comfort level with a changing technological landscape.
We are not focussed on chasing down individuals for experimenting with the work of others. Our concern is less with artists having fun, rather with the industrial scale usage of artist training data.
On a smaller level, we do see a lot of potential in offering official channels to experiment with artist training data and establish channels for provenance and remuneration, similar to what we pioneered with Holly+.
While legal proceedings that determine the fate of copyrighted works are out of our hands, what is within our grasp is the ability to conceive of a new era of consenting interactions around artist data.
A new era offers us the opportunity to reconfigure how we treat IP! We believe that the best path forward is to offer individual artists tools to manage their style and likenesses, and determine their own comfort level with a changing technological landscape.
We are not focussed on chasing down individuals for experimenting with the work of others. Our concern is less with artists having fun, rather with the industrial scale usage of artist training data.
On a smaller level, we do see a lot of potential in offering official channels to experiment with artist training data and establish channels for provenance and remuneration, similar to what we pioneered with Holly+.
How does an artist get approved to use Spawning tools?
Once you have applied for beta access, a human will curate and verify access for artists on a case by case basis. We seek to ensure that data is labelled accurately.
Do you distinguish between artists using traditional tools and promptists using AI tools?
Yes, although we have a tool in the works for promptists too. We feel that verifying and labelling good artist data expands the palette for promptists. As a result of our efforts there will be more to prompt and experiment with. We respect promptists as artists in their own right.
©2023 Spawning, Inc.
Building tools to manage your AI identity.