The New York Times has taken legal action against OpenAI and Microsoft, alleging copyright infringement in their utilization of Times content to train generative AI and large-language models like ChatGPT and Bing AI. The lawsuit, filed in a Manhattan federal court, claims these tech entities placed significant emphasis on Times material without authorization or compensation, potentially amounting to billions of dollars in damages.
The lawsuit highlights the unlawful use of the Times’ copyrighted articles, including news, investigations, opinion pieces, reviews, and more, to develop AI products. It asserts that this exploitation poses a threat to the Times’ ability to deliver its journalistic services, which the lawsuit argues are vital for democracy in today’s society.
This legal move by the Times follows a pattern of similar cases, indicating growing concerns among content creators about AI systems utilizing their work without proper permission or acknowledgment. The lawsuit also delves into the issue of AI-generated misinformation that could harm the Times’ reputation, citing instances of misattributed content on Microsoft’s Bing Chat.
Although the lawsuit does not specify a monetary claim, it urges accountability from OpenAI (valued at $80bn) and Microsoft (valued at $2.8tn) for substantial statutory and actual damages. Additionally, it demands the destruction of any AI models and training data incorporating copyrighted Times material.
Amid rising tensions over the use of published content to train AI systems, OpenAI has sought partnerships to address these concerns. One such collaboration with Axel Springer, a German multinational mass media company, aims to integrate recent and credible content into ChatGPT, emphasizing transparency and attribution to content creators.
While OpenAI has expressed surprise and disappointment at the lawsuit, claiming constructive ongoing conversations with the Times, the legal battle underscores the growing conflict between information providers and tech operators using their content without explicit authorization.