Thanks for the link to the paper, I'll comment on what I find to be relevant in it in case it is of use to others. There is a part re Sega v Accolade that brin

Thanks for the link to the paper, I'll comment on what I find to be relevant in it in case it is of use to others. There is a part re Sega v Accolade that brings up the interesting sub-topic of "intermediate copying", which is legal, but then says it's legal because it's “the only way to gain access to the ideas and functional elements embodied in a copyrighted computer program.” But wholesale scanning of entire works is not the only way to gain access to the ideas of books. It's also possible to purchase copies from the publishers. Perhaps a bit more relevant- "Accolade was not using Sega’s games for their protected expressive content, but simply to extract some unprotected, functional, nonexpressive information contained within them" I think that would be debatable in certain machine learning projects. Perhaps much of it would fall under that category, but a large part of the point would in fact be to extract the expressive information contained in them. What I'm trying to get at is related- Google (and other machine learning giants) is making an effort to extract the expressive content of these works. That's different from what we've seen in the past. To me, at least, it feel like it's worth asking if it's right that the people who produced that expressive content should not be compensated for it.