Recently, Microsoft has been sued by several renowned authors, who accuse the company of using a large number of pirated books without permission in training its artificial intelligence model Megatron. The suing authors include Kai Bird, Jia Tolentino, and Daniel Okrent, who filed a lawsuit in the U.S. District Court for the Southern District of New York, claiming that Microsoft used pirated electronic versions of their works for AI training, enabling the AI to generate responses based on users' text instructions.
The core dispute in this lawsuit is whether the use of copyrighted materials during AI training constitutes copyright infringement. In addition to Microsoft, tech companies such as Meta, Anthropic, and OpenAI are also named as defendants. These companies claim that using copyrighted content falls under "fair use," stating that their purpose is to create transformative new products. However, the complaint points out that the data set used by Microsoft contains nearly 200,000 pirated books, raising widespread questions about its legality and ethics.
Notably, the complaint mentions that Microsoft's Megatron model not only relies on the labor of thousands of creators but can also generate texts that mimic the grammar, style, and themes of the original works. This has made creators feel that their works are being unfairly exploited and has led to calls for legal protection.
On the day before this case was filed, a federal judge in California ruled that Anthropic's use of authors' works in training AI could be considered "fair use," but if the content used was pirated, it might still be held legally liable. This ruling is seen as the first legal judgment in the United States regarding whether generative AI can use copyrighted works without authorization.
The authors hope the court will issue an injunction to stop Microsoft from continuing its infringing activities and demand that it pay statutory damages of up to $150,000 per infringed work. This lawsuit not only concerns the rights of individual creators but also sparks deeper reflection on the balance between AI technological development and copyright protection.
Key Points:
📚 Microsoft is sued by multiple authors for using pirated books to train the AI model Megatron.
⚖️ The core of the lawsuit is whether AI training infringes on copyrights, involving several tech companies including Meta and Anthropic.
💰 The authors hope the court will issue an injunction and demand compensation, with a maximum of $150,000, to protect the rights of creators.
