New York Times-Led Publishers Ask Court to Sanction OpenAI in U.S. Copyright Lawsuit

OpenAI slammed by media giants over ‘improper handling’ of evidence in copyright fight The New York Times-led publishers' group asked a federal court to sanction the AI firm in a high-profile legal dispute. In a highly charged legal spat over whether OpenAI improperly handled evidence, the media giants who sued over use of copyrighted works in its AI are now pushing a US court to impose sanctions. A coalition of news outlets, including the The New York Times, claimed the artificial intelligence giant “failed to properly preserve potentially important evidence” in a copyright infringement suit over content used to train AI models, according to a court filing made public on Thursday.

The publishers are locked in a battle with OpenAI, arguing that it used their works without permission or proper compensation to develop its large language models (LLMs).

They are seeking to require AI firms like OpenAI to obtain licenses to use copyrighted material in the training process. OpenAI has maintained that it operates within existing copyright law and does not infringe the law by not licensing material used to train its LLMs, which it argued doesn’t actually reproduce copyrighted material in output. The latest filing from the publishers’ side represents a new phase in a landmark legal case that could affect how AI companies, creators, and copyright laws interact. The publishers contend that OpenAI did not properly preserve information that could have been vital in substantiating their case after the lawsuit began.

OpenAI disputes the allegations.

This legal battle is one of the most high-profile cases involving the generative AI industry and the application of intellectual property to cutting-edge technology. Copyright holders have been bringing similar lawsuits against other AI firms, raising the question of how to regulate and compensate for content used to fuel rapidly developing AI technologies. The core of the suit centers on the publishers’ claim that OpenAI's practice of using their published works as data to train its LLMs is copyright infringement.

Copyright law generally requires obtaining permission to use copyrighted works for such purposes. However, OpenAI’s legal team argues that the training of their models is a transformative process and therefore constitutes fair use, falling within the permissible uses allowed under US law. The recent move to request sanctions is a procedural dispute that highlights the escalating tensions between the media industry and AI developers.

The publishers argue that a failure to preserve evidence by OpenAI is prejudicial to their ability to properly prosecute the lawsuit.

They said that even if the shortcomings were unintentional, the impact on their case necessitates court intervention. In their response, OpenAI rejected the publishers’ characterization of events, asserting that there are no grounds for sanctions and that they have met all their legal obligations during discovery. The company is expected to respond formally to the publishers’ motion, and this issue will be taken up by the court as the case continues. This dispute, among other lawsuits from authors, musicians, and other copyright owners, is critically important for establishing a framework for the AI era.

The rulings on these cases will likely determine the licensing landscape for creative content and could influence the future direction of AI development and its business models.

It has become clear that while legal frameworks have existed for decades to govern intellectual property, they are being severely tested by the ability of AI to analyze, learn from, and generate content at an unprecedented scale.