|
|
|
|
|
by throwaway82388
1400 days ago
|
|
Perhaps I’m misunderstanding your argument, but my counterexample would be: if a human digital artist transformed a Getty image, resulting a fantastical, never-before-seen result, using software like Photoshop, that use would be no more defensible. If anything, the vast scale at which this occurs in AI makes it worse. |
|
For example:
Andy Warhol died in 1987, 35 years ago. One of his 'Prince' collages dating to the early 80s used another photographer's photo, without permission. In 2019, one federal judge ruled that was not infringement. An appeals judge then said it was.
The Supreme Court has decided to take the case.
The US Copyright Office & Department of Justice agree with the photographer in briefs filed with the court... but the mere fact the Supreme Court took the case indicates they think there might be issues with the appeals court ruling. They might agree with the original judge!
Oral arguments come this October. See:
https://www.reuters.com/legal/litigation/us-backs-photograph...
So, when all the (possible) disputes over AI-training-on-copyrighted-images resolve – maybe in the 2030s or 2040s? – what will the laws say, & courts decide? It'll depend a lot on other specifics, & reasoning, that may not be evident now.