Training in the US has, in fact, been decided (at least to the extent that anything has currently been decided). Bartz vs Anthropic specifically ruled training an AI model on legally owned copyrighted material is sufficiently transformative[1]:
This order grants summary judgment for Anthropic that the training use was a fair use.
And, it grants that the print-to-digital format change was a fair use for a different reason. But it
denies summary judgment for Anthropic that the pirated library copies must be treated as
training copies.
The document you linked was written a month before the Bartz decision was reached. It's also worth noting even the document you linked says this in its conclusion:
Various uses of copyrighted works in AI training are likely to be transformative. The
extent to which they are fair, however, will depend on what works were used, from what
source, for what purpose, and with what controls on the outputs—all of which can affect the
market.
[1]:
https://copyrightalliance.org/wp-content/uploads/2025/06/Bar...