Microsoft argues Supreme Court’s VCR ruling should doom NYT’s OpenAI lawsuit

In a legal battle reminiscent of the 1980s VCR controversy, Microsoft has urged a federal court to dismiss part of The New York Times’ copyright lawsuit against itself and OpenAI. The tech giant drew parallels between the movie industry’s attempts to halt VCR technology and the NYT lawsuit, alleging that both center on copyright infringement claims. Microsoft’s filing, referencing the iconic 1982 quote by MPAA President Jack Valenti, likened the NYT’s arguments to those rejected by the Supreme Court in the Sony Corporation of America v. Universal City Studios case.

The crux of the issue lies in the NYT’s accusations against OpenAI and Microsoft for using large language models (LLMs) developed by infringing on millions of NYT copyrighted works. Microsoft refuted these claims, asserting that liability for users’ copyright infringement cannot be pinned solely on offering multi-use products, drawing parallels between LLMs and previous transformative technologies such as the VCR, player piano, and the Internet.

The lawsuit takes a contentious turn as OpenAI accuses The New York Times of manipulating their AI tools to generate content matching NYT’s works, a claim vehemently contested by Microsoft. The legal battle intensifies with the NYT alleging direct, contributory, and vicarious copyright infringement against Microsoft and OpenAI in the training and use of LLMs. However, Microsoft’s motion seeks to dismiss claims of contributory infringement while preparing to contest direct and vicarious infringement allegations under a fair-use defense.

Despite the disparity between using AI models and physical media like tapes, Microsoft underscores the striking similarity in legal challenges stemming from the design or distribution of products capable of lawful use. Emphasizing the lawful applications of LLMs, Microsoft pushes back against the NYT’s claims of aiding infringement, highlighting the lack of real-world instances of copyright violations due to the use of GPT-based products.