Court Finds That Using Copyrighted Material as AI Training Data Is Not Fair Use
A federal court ruled that utilizing copyrighted works to train artificial intelligence models does not qualify as fair use under current law. This decision establishes a significant legal precedent for technology companies, which have previously argued that ingesting public data to develop large language models is transformative and permissible. The outcome creates new liability risks for developers and potentially mandates that AI firms secure licensing agreements or obtain permission from creators before training future iterations of their software.
Covered by 1 source
- FFuturism↗Oct 6