4 ms·

Google “expert witness”! Courts are also known to hire their own experts who mediated between the experts on either side. Also, those emails seem very likely t

Google “expert witness”! Courts are also known to hire their own experts who mediated between the experts on either side. Also, those emails seem very likely to be ordered to be produced during discovery. This thing could really go either way at this point but I feel like Stability has the upper hand. Imagine training a model without any of the plaintiffs images, then using that side by side with the model that does. This could then be used to show the jury that those individual works are of no importance to the system if the images are of the same quality. They will probably argue that the individual expressions of each work are not copied, rather the abstract ideas of two-dimensional representations present across any and all images. Expect lots of side by side pictures as Exhibits from both side! Grandma and her fellows have to weigh in on this one! This is a fun one!

kranke155 3y ago
Stable Diffusion has been known to make virtually identical copies of the images it was "trained" on, afaik. If the images are REALLY of no importance, they wouldn't have been used anyway.
- williamcotton 3y ago
  And then the defense reminds the court that copyright is about specific works. Clearing the specific intent to copy a specific work is pretty key if you want to argue idea/expression or fair use or whatever their strategy may be. And yes, the nearly identical copies will definitely be presented to the jury.
- michaelbrave 3y ago
  that's a gross oversimplification and an argument that was made in bad faith that has now spread like a meme. The gist was if you overtrain a model, and try to recreate an exact image by prompting very similar things, and run it several thousand times, then you can recreate an image. If you seriously crank up the overtraining it's even easier. But normal use of the models do not just pump out recreations of training data.
- WithinReason 3y ago
  That never happened afaik

kranke155 3y ago

Stable Diffusion has been known to make virtually identical copies of the images it was "trained" on, afaik. If the images are REALLY of no importance, they wouldn't have been used anyway.

williamcotton 3y ago
And then the defense reminds the court that copyright is about specific works. Clearing the specific intent to copy a specific work is pretty key if you want to argue idea/expression or fair use or whatever their strategy may be. And yes, the nearly identical copies will definitely be presented to the jury.
michaelbrave 3y ago
that's a gross oversimplification and an argument that was made in bad faith that has now spread like a meme. The gist was if you overtrain a model, and try to recreate an exact image by prompting very similar things, and run it several thousand times, then you can recreate an image. If you seriously crank up the overtraining it's even easier. But normal use of the models do not just pump out recreations of training data.
WithinReason 3y ago
That never happened afaik