I read a NYT article and publish a summary of facts that I learned: totally legit.
Train a model on NYT text that outputs a summary of facts that it learned: OMG literally murder.
Train a model on NYT text that outputs a summary of facts that it learned: OMG literally murder.
I read a NYT article and publish an exact copy of that article on my website: copyright infringement.
Train a model on NYT text and it outputs an exact copy of that text: also copyright infringement.
Probably not until they pay him a hefty copyright fee.
Determining whether a work violates a copyright requires holistic consideration of the similarity of the work to the copyrighted material, the purpose of the work, and the work’s impact on the copyright holder.
There is not an algorithm for this, cases are decided on by people.
There are algorithms that could detect obvious violations of copyright, such as the one you suggest which looks for exact matches to copyrighted material. However, there are many potential outputs, or patterns of output, which would be copyright violation and would not be caught by this trivial test.
What does that mean?
Look up "substantial non-infringing use" and this little court case:
https://en.wikipedia.org/wiki/Sony_Corp._of_America_v._Unive....
Now spend a few million on lawyers and roll your dice.
If you can't copyright AI-generated pieces, then why would fair use apply to LLMs?
Is it? Can you quote relevant legislation or case law?
Also remember copyright laws was not there in the first place.