More important than an opinion of a court is the opinion of lawmakers. After all, copyright laws have a purpose and AI subverts that purpose. Lawmakers will have to decide whether the benefit of AI to society is greater than the benefit of having works published. We will see laws adapted accordingly.
> After all, copyright laws have a purpose and AI subverts that purpose.
While it may differ in other countries, in the United States the purpose of intellectual property laws, as expressed in the Constitution, is to "To promote the Progress of Science and useful Arts by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries".
Enriching the copyright holders (on the rare occasions that actually occurs) is a secondary consequence, not the prime purpose.
Does AI "subvert" "promoting the Progress of Science and useful Arts"? I don't think so. Quite the contrary... I think it advances the progress of science and the useful arts, if anything.
It's pretty well established that a description of a copyrighted work is not protected, even an extremely-detailed description (see, for example, the way that Phoenix assigned one team to write an extremely-detailed description of the IBM PC BIOS, then gave that description to a second clean-room team that hadn't seen any of the actual source code. The second team then produced a clone of the BIOS that could be sold without paying IBM anything).
The data stored in these models seems more like a "description" rather than a "copy" to me -- though, of course, there's no guessing what a court or legislature will decide.
> Does AI "subvert" "promoting the Progress of Science and useful Arts"? I don't think so. Quite the contrary... I think it advances the progress of science and the useful arts, if anything
It does subvert that purpose to the extent that it makes some people no longer willing to share their works. The entire purpose of copyright is to encourage the sharing of works.
