The ruling holds that training on publicly accessible text can be fair use where the output does not substitute for the original. It also finds that ignoring machine-readable opt-outs weighs against the defendant. Lawyers describe it as a roadmap rather than a blank cheque.
What happens next
Expect widespread adoption of opt-out signals, licensing deals for premium content and appeals. Developers are reviewing their crawling practices. Publishers are updating their robots files.
Key takeaways
- Fair use upheld with substitution test
- Opt-out signals carry legal weight
- Appeals are expected
This story is part of our ongoing coverage of policy & privacy. We will update it as more details emerge.
