The TechCrunch AI report highlights a critical nuance often lost in the broader AI copyright debate: context matters. While recent California rulings have leaned toward a fair use interpretation for training data scraped from the public web, this case involves a more structured, permission-based relationship. Google's alleged use of books provided under specific, limited licenses for search—not for model training—could be viewed by a court as a breach of that understanding, regardless of the abstract fair use question.
The outcome here may hinge less on copyright law's general provisions and more on the specifics of contractual intent and the scope of implied consent, setting a potential precedent for how companies handle data provided under one set of rules for a different, more lucrative purpose. This moves the debate from 'is scraping legal?' to 'what did you promise?', a much thornier question for the industry.