Credible argument? You call baiting and entrapment credible? Those articles don't pop up on their own from chatGPT, it takes effort. You need to know the first phrase and the title to prime it. And then it only works once in a while, not every time. They even complained in the lawsuit that the cases where the model doesn't regurgitate it hallucinates articles as from them and they don't like that either.
NYT itself is full of fluff and derivative news, and suing for expired news that only have historical value now. It's just flexing their power to corner AI.
We can't copyright everything. Some things are free: letters, words, sometimes common phrases, also, ideas. An AI is entitled to learn everything at the level of ideas, even from copyrighted works. They just need to paraphrase or rewrite as Q&A the copyrighted material before training to avoid identical replication.
Copyright is already too big as it is, don't allow them to make ideas protected as well, it will have a chilling effect on us too, not just AI models.
Baiting and entrapment? Please. They used the API the way it was designed. The next thing, you'll be arguing that hot women in nice dresses are somehow "entrapping" their rapists.
If anything, their examples illustrate the kind of knowledge that can only be found in the NYT's original reporting. They focused on some events that only their newspaper covered well and the AI responded by plagiarizing the NYT's coverage, something that's part of its inherent design.
And I'm not talking about copyrighting everything. Nor is the NYT. Consider the opposite case where they have no way to force readers to pay them for their hard work. Is that the kind of world you want to live within?
I'm very happy to reward the artists and writers who make creations. I don't see anything wrong with giving them what anti-copyright people call a "monopoly." Why? Because I know that it pushes other people to create their own new knowledge for society and not skate through by simply plagiarizing.
NYT itself is full of fluff and derivative news, and suing for expired news that only have historical value now. It's just flexing their power to corner AI.
We can't copyright everything. Some things are free: letters, words, sometimes common phrases, also, ideas. An AI is entitled to learn everything at the level of ideas, even from copyrighted works. They just need to paraphrase or rewrite as Q&A the copyrighted material before training to avoid identical replication.
Copyright is already too big as it is, don't allow them to make ideas protected as well, it will have a chilling effect on us too, not just AI models.