US government backs OpenAI in copyright case over AI training data

Key facts

• The U.S. government filed a brief supporting OpenAI in its copyright dispute with The New York Times.
• The brief argues that training AI models on copyrighted material constitutes fair use.
• The government cited national security and American prosperity as reasons to support AI development.
• The New York Times accuses OpenAI and Microsoft of using millions of articles without permission to train AI models.
• The government stated that constraining AI development would hinder scientific progress and economic mobility.

The U.S. government has entered the legal fray over artificial intelligence training data, filing a brief in support of OpenAI in its copyright dispute with The New York Times and other news outlets. The brief, submitted to a Manhattan federal court, represents the government's first official stance on the numerous lawsuits filed by copyright holders against AI companies.

The government's filing argues that the use of copyrighted material to train large language models (LLMs) constitutes fair use under copyright law. This position aligns with arguments made by tech companies fighting these claims, which contend that AI training is a transformative process.

The New York Times initially filed its lawsuit in 2023, alleging that OpenAI and its major investor, Microsoft, illicitly used millions of newspaper articles to train ChatGPT. This case is part of a broader wave of litigation from authors, publishers, and media companies concerned about the unauthorized use of their content for AI development.

In its brief, the U.S. government emphasized its interest in rejecting arguments that AI training violates copyright law, citing potential impacts on scientific advancement and national security. The filing also stated that hindering LLM development could thwart creative and scientific progress, thereby impacting American prosperity and economic mobility.