The New York Times is suing ChatGPT's owner, OpenAI, for copyright infringement for training the system. The lawsuit, which also includes Microsoft as a defendant, claims that the companies should be held liable for "billions of dollars" in damages.
ChatGPT and other large-scale language models (LLMs) "learn" by analyzing huge amounts of data, often drawn from online resources. The lawsuit alleges that "millions" of articles published by the New York Times were used without its permission to improve ChatGPT, which the newspaper believes has led to competition with it as a source of trusted information.
It is alleged that ChatGPT, when queried about current events, sometimes generates "verbatim excerpts" from New York Times articles that cannot be accessed without a subscription. This, according to the New York Times, leads to a loss of subscription revenue and ad clicks.
The lawsuit reveals that the New York Times unsuccessfully tried to reach a "settlement" over the copyright issue with Microsoft and OpenAI in April. The legal wrangling comes after internal problems at OpenAI and a series of lawsuits, including the ouster and reinstatement of the firm's CEO and a case over the use of code to train artificial intelligence.
None of the lawsuits have been resolved to date.