Anthropic Achieves Major AI Copyright Victory, Yet Braces for Trial Over Piracy Allegations

Anthropic has achieved a significant legal win regarding its use of copyrighted works to train AI models, a case that may influence other AI copyright lawsuits in the U.S. A court ruled that Anthropic’s use falls under the “fair use” doctrine, which allows for certain unauthorized uses of copyrighted material. Senior District Judge William Alsup determined that the training method was “transformative,” meaning it creates something new rather than replicating the original work. However, this victory comes with limitations; the court allowed authors to pursue claims regarding the pirated materials Anthropic initially collected—over seven million unauthorized copies. Although Anthropic now trains on purchased works, it had maintained a library of these pirated copies. The lawsuit, filed by authors in August 2024, will continue to trial regarding these copyright violations. Anthropic expressed satisfaction with the ruling, which aligns with copyright’s intent to foster creativity and progress.

Source link