In a significant development for legal professionals tracking intellectual property disputes in the realm of artificial intelligence, Microsoft Corp. is facing a lawsuit in the United States District Court for the Southern District of New York. The tech behemoth has been accused of utilizing pirated copyrighted material to train its AI models, particularly its Megatron-Turing Natural Language Generation Model (MT-NLG), without obtaining the necessary consent from content creators. The complaint was filed by a coalition of writers, which includes Pulitzer Prize winner Kai Bird, and is supported by prominent law firms such as Lieff Cabraser Heimann & Bernstein, Cowan Debaets Abrahams & Sheppard, and Susman Godfrey.
Microsoft’s legal challenges arise in the wake of two pivotal legal rulings in cases involving other AI developers, where similar copyright infringement claims were made. Both Meta Platforms Inc. and AI firm Anthropic were implicated in lawsuits for their use of pirated literature to train AI models, including using a dataset known as “Books3,” which allegedly comprises around 200,000 pirated books.
The suit against Microsoft, reported by Law.com, demands injunctive relief and statutory damages potentially reaching $150,000 per infringed work. This case presents another complex intersection of AI technology and copyright law, posing substantial challenges and implications for AI developers and legal practitioners alike.