Microsoft Files Motion to Dismiss NYT’s Infringement Lawsuit, Calling It “Unsubstantiated” Using Deceptive Prompts

By Moon Harper

Mar 06, 2024 04:54 AM EST

Microsoft is accusing The New York Times of "unsubstantiated" allegations in a lawsuit filed in December against OpenAI, which could significantly impact the future of generative artificial intelligence.

In a recent motion to dismiss filing on Monday, Microsoft criticized The New York Times for portraying OpenAI's ChatGPT as a threat to the news industry with a "doomsday futurology" narrative. It likened the lawsuit to Hollywood's past resistance to VCR technology in the 1970s, enabling users to record television programs.

Growing Tension Over Large Language Models and Media Sector

Attorneys representing Microsoft stated that The New York Times is leveraging its influence to contest the latest significant technological advancement, the Large Language Model. Microsoft is OpenAI's primary investor, having injected approximately thirteen billion dollars into the company.

The submission represents the newest development in the ongoing conflict between OpenAI and the media sector, which is growing apprehensive about AI models being trained on extensive content accumulated over decades. The Times, in its legal action, alleged OpenAI and Microsoft of copyright infringement and misusing the newspaper's intellectual property while training LLMs.

OpenAI's Motion to Dismiss NYT's Lawsuit

OpenAI has petitioned a judge to dismiss sections of the lawsuit filed by the Times, asserting that the publisher engaged in unethical behavior by paying someone to hack OpenAI's products, including ChatGPT, to fabricate 100 instances of copyright infringement as evidence. OpenAI contended that the Times needed tens of thousands of attempts to create these results and used "deceptive prompts" that directly violated OpenAI's terms of use.

In its recent submission, Microsoft's legal team contends that the content utilized to train LLMs does not replace the market for the works but instead instructs the models' language. A representative from The Times did not promptly respond to CNBC's inquiry for a comment.

Since its release to the public in late 2022, OpenAI's ChatGPT has propelled the company to become one of the most sought-after startups globally, boasting a valuation that exceeds eighty billion dollars.

OpenAI has recently stated that training advanced AI models without using copyrighted materials is "impossible." In a filing last month in the U.K. House of Lords, OpenAI explained that copyright law covers a wide range of human expression, including blog posts, photographs, forum posts, software code snippets, and government documents, making it impossible to train leading AI models without utilizing copyrighted content.

During an event in Davos, Switzerland, in January, OpenAI CEO Sam Altman expressed his surprise at The New York Times' lawsuit, stating that OpenAI's models did not require training using the publisher's data.

At an event organized by Bloomberg in Davos, OpenAI CEO Sam Altman stated that they do not need to train on The New York Times' data. He emphasized that people often misunderstand this point, clarifying that any single training source does not significantly impact their models that much.

OpenAI has entered agreements with Axel Springer, the German media conglomerate behind Business Insider and Morning Brew, among other outlets. It is reportedly in discussions with CNN, Fox Corp., and Time for licensing their content.