Dr. Susan L. Karamanian
A recent lawsuit filed in a US federal court by The New York Times Company against defendants Microsoft Corporation and OpenAI entities alleged that the defendants’ generative artificial intelligence (GenAI) tools unauthorizedly used “millions” of The Times’s original work. The copied content, according to The Times, was costly to create and reflected tremendous investigative skill and professional judgment.
The lawsuit sets out how Microsoft’s Bing Chat (later Co-Pilot) and OpenAI’s ChatGPT copied content – some of it under a paywall - and even incorrectly linked The Times to information it had not published, or what is referred to as the practice of “hallucination.” Further, the lawsuit alleges that applicable algorithms, large-language models (LLMs), were “trained” using content from The Times.
For many years technological innovation, reflected in digitalization, has run up against copyright law, which gives exclusive rights to creators. One need only recall Napster, the music download site that has been shuttered. Legal sagas between creators and YouTube are still ongoing. Like other cases, The Times’s lawsuit seeks a court order compensating the creator whose work is copied or forms the basis of content and an injunction to stop defendants from copying in the future.
The defendants claim they have the right under the “fair use” doctrine, set out in 17 USC. section 107, to use a copyrighted work “for purposes such as criticism, comment, news reporting, teaching…scholarship, or research.” Fair use, however, is not absolute and the following statutory factors also set out in section 107 shape its application: the “purpose and character of the use” including whether it is commercial or nonprofit educational; the type of work being copied; the amount copied relative to the entire work; and the “effect of the use upon the potential market for or value of the copyrighted work.”
Even if The Times and the defendants can craft a settlement, for example based on The Times’s providing a license to defendants based on defendants’ payment of royalties, other creators may be next in line to sue. Some creators, however, may not pursue litigation as it is burdensome; or they may appreciate the publicity having their work in some fashion linked to GenAI content.
Yet The Times’s lawsuit is eye-opening beyond its generic traditional copying allegations given its revelation about the Microsoft-OpenAI relationship, the shifting of OpenAI from a non-profit to a commercial endeavor, and the nature of the LLMs. The Times admits limited knowledge about the operations of newer LLMs, and if the case goes to the “discovery” phase, it may gain this information. If “fair use” is on the table, then surely the entirety of the algorithms and training sources should be disclosed and publicly disseminated given their value to research.
The thoughts and views expressed are the author’s own and do not necessarily reflect an official University stance.
