News for hackers. Stuff that matters.
Re:Artists Lose First Round of AI Art Copyright Infringement... by kranke155 on Tuesday October 31, 2023 @08:38AM (#38083861)
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.

Re:Artists Lose First Round of AI Art Copyright Infringement... by williamcotton on Tuesday October 31, 2023 @08:43AM (#38083915)
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.

Re:Artists Lose First Round of AI Art Copyright Infringement... by michaelbrave on Tuesday October 31, 2023 @08:47AM (#38083978)
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.

Re:Artists Lose First Round of AI Art Copyright Infringement... by WithinReason on Tuesday October 31, 2023 @09:36AM (#38084594)
That never happened afaik