Machine learning requires the copying of extraordinary amounts of copyrighted material. That copying should generally be permitted. Most ML systems copy works not to consume the expression copyright law protects, but to get access to the facts or structures copyright law dedicates to the public. Understanding this as fair learning can help ensure we can train ML systems without interference from the law. But the idea of fair learning doesn’t just matter for robots. It can help us resolve a number of troubling copyright cases involving humans too. And it reminds us that fair use is about more than just transforming copyrighted works into new works. It’s about preserving our ability to create, share, and build upon new ideas. In other words, it’s about preserving the ability to learn—whether the entity doing the learning is a person or a robot
Mark A. Lemley and Bryan Casey
Texas Law Review
March 2021
External Link
Loading...