Wikipedia urges middle ground in New York Times vs OpenAI lawsuit

A middle ground for fair use

The Wikimedia Foundation has officially entered the legal battle between the New York Times and the AI giants OpenAI and Microsoft. By filing an amicus brief, the organization is asking the court to find a balanced interpretation of fair use law. This legal concept determines when someone can use copyrighted material without permission from the original owner.

The lawsuit centers on whether OpenAI and Microsoft infringed on copyrights by using Times articles to train their models. It also examines whether these models are allowed to reproduce specific portions of those articles in their outputs. Wikipedia's goal is to ensure the court does not swing too far in either direction of the debate.

As A.V. Club reported, the foundation wants to avoid a result that favors one extreme over the other. They believe that a nuanced ruling is necessary to protect both the creators of content and the developers of new technologies. The organization is positioning itself as a mediator in a high-stakes fight over the future of digital information.

The risks of extreme rulings

The Wikimedia Foundation warns that a ruling favoring a very restrictive version of fair use could be damaging. They argue that such a move would inhibit technological innovation and limit the availability of tools that benefit open-source projects. If developers cannot use data to train models, it could slow down the progress of many public-interest technologies.

Conversely, the foundation argues that a ruling in favor of blanket fair use could result in a more restricted internet. In this scenario, knowledge online might be weakened because there would be a lack of available secondary source materials. They fear that if everything is automatically considered fair use, the economic and legal incentives to create and share information could disappear.

The organization is specifically looking for a distinction between training and output. They suggest that while large language models should be allowed to train on copyrighted materials, those materials should not be used in retrieval-augmented generation. This is a process where a chatbot pulls in specific external data to provide a more accurate or current answer to a user's prompt.

The sustainability of free knowledge

Stephen LaPorte, the General Counsel for the Wikimedia Foundation, expressed concern about the long-term health of the internet. He stated that a decision that does not handle fair use rules with precision could jeopardize the continued existence of free information. He believes that the current legal uncertainty could ultimately stifle the development of open-source technologies.

The foundation's position is rooted in the idea that access to information must be preserved for everyone. They argue that the current situation impacts how people interact with information on a global scale. If the legal framework is too rigid, it may prevent the creation of tools that help people find and understand the world around them.

The brief highlights a delicate balance between the rights of publishers and the needs of the public. By advocating for a middle ground, Wikipedia hopes to create a legal precedent that allows for progress without erasing the rights of the people who do the work. They want to ensure that the internet remains an open space for learning and discovery.

Wikipedia's complex AI relationship

Wikipedia's involvement in this case is notable because the site has a complicated relationship with artificial intelligence. These models have already been training on the hard work of volunteer contributors who maintain the encyclopedia. At the same time, these tools may be pulling traffic away from the website, reducing its reach and visibility.

There is also a concern regarding how AI handles the specific licenses used by Wikipedia. The site uses Creative Commons licenses, but AI models may reproduce copyrighted material without adhering to the specific terms of those licenses. This creates a situation where the site's content is used to build products that do not follow the rules the site itself must obey.

Despite these concerns, Wikipedia has acknowledged that AI tools have provided some benefits to its community. Editors have used these technologies to help detect plagiarism and identify instances of vandalism on the site. This dual reality of AI as both a threat to traffic and a tool for maintenance defines the foundation's cautious approach to the lawsuit.

Current policies and next steps

The Wikimedia Foundation has already taken proactive steps to manage AI's influence on its platform. Earlier this year, the organization chose to prohibit AI-generated content from appearing on its pages. This move was intended to preserve the human element of the encyclopedia and ensure the quality of the information provided to readers.

While they have restricted AI-generated content, they continue to monitor how the technology interacts with their data. The outcome of the New York Times lawsuit will provide a significant legal framework for how these interactions are handled in the future. The court's decision will likely set the standard for how other open-source projects interact with copyrighted data.

The foundation's amicus brief serves as a signal of their priorities in the coming years. They are looking for a legal environment that permits the growth of open-source tools while protecting the integrity of the source material. The next phase of this legal battle will determine if a middle ground is actually possible in the eyes of the law.