As it stands, your article is highly misleading, with that "uncertainty" sharing an acronym with "fear" and "doubt". I'd prefer to assume that ain't intentional.
It might be easier to take them from NetBSD; it wouldn't introduce the GPL licensing issue, and courtesy of their rump kernel system they're actually kind of designed for it.
In general, now that the pump has been fully primed for capital to flow into developing "AI", I do not see how copyright law is going to make much of a dent in that trend. Nor do I see how "AI" companies are going to make a dent in copyright law for anyone but themselves. I foresee large "AI" companies being essentially unbound on training over small-owner copyrighted works, upstart "AI" companies needing to pay into a hefty protection racket, and individuals still bound by imaginary property laws whether directly (old fashioned piracy) or when using common genAI (sorry Dave, I can't do that).
I just ran into a situation where ChatGPT refused to quote me the relevant bit of the electrical code for my state (supposedly binding law), because those laws were created by wholesale importing the "National Electrical Code" which is copyrighted. At best, the situation is an open legal question. And yet de facto there is still a restriction that prevents me from using the tool to engage with the law in good faith.
I thought I made that pretty clear when I wrote in my original comment that "[i]ntellectual property and the enforcement thereof is in and of itself a Torment Nexus."
> In general, now that the pump has been fully primed for capital to flow into developing "AI", I do not see how copyright law is going to make much of a dent in that trend. Nor do I see how "AI" companies are going to make a dent in copyright law for anyone but themselves.
"AI" exists outside of the various corporations hosting LLMs on The Cloud™. The corporate-hosted LLMs get undue emphasis largely as yet another result of the Torment Nexus that is intellectual property.