The Wikimedia Foundation, the nonprofit entity operating Wikipedia, has joined the legal battle involving The New York Times, OpenAI and Microsoft. The organization has asked a US court to strike a measured balance when determining whether artificial intelligence developers have the right to utilize copyrighted content for training models and producing outputs.
In connection with the ongoing New York copyright lawsuit, Wikimedia submitted an amicus brief, or friend-of-the-court filing. The nonprofit contends that a decision that swings too far toward restriction or permission could harm the open internet and restrict public access to trustworthy information.
Wikimedia Proposes Two Different Rules for AI
Rather than fully aligning with either publishers or AI developers, Wikimedia asks the court to draw a clear line between training AI models and deploying online material to craft chatbot responses.
According to the foundation, utilizing copyrighted texts to train large language models (LLMs) typically amounts to fair use whenever the resulting artificial intelligence system generates novel and transformative content distinct from the source material.
The organization holds a distinct view regarding retrieval-augmented generation (RAG), a system enabling chatbots to pull live data from web pages to address user inquiries.
Wikimedia maintained that whenever AI products duplicate existing stories via RAG, developers ought to honor the licensing terms of the original publishers instead of simply relying on a fair use defense.
Also Read: OpenAI and Copyright: The Battle Between AI and Creative Rights
Why Wikipedia Wants Proper Attribution
Wikipedia depends on a community of thousands of volunteer authors who draft and update entries, resulting in a repository exceeding 67 million articles spanning 300 languages. Because this database operates under the Creative Commons CC BY-SA 4.0 license, reuse and modification are permitted provided that creators receive appropriate credit and derivative works carry matching licenses.
According to Wikimedia, AI firms must adhere to these terms when pulling and displaying Wikipedia material. If attribution is absent, conversational bots might leverage data while ignoring the individuals who authored it.
Also Read: No Copyright Breach, Delhi High Court Backs OpenAI Training
Copyright Ruling Could Shape AI’s Future
Brought forward in 2023 by The New York Times, the initial complaint accuses OpenAI and Microsoft of copyright violations through the unauthorized harvesting of journalistic works. Both tech companies maintain that leveraging public internet data for AI research constitutes fair use.
Wikimedia cautions that permitting unconstrained reproduction risks undermining publishers and external content providers. Conversely, placing excessive legal hurdles on AI training could impede progress and hinder open-source advancements. The upcoming judicial ruling is poised to impact how artificial intelligence businesses gather, analyze, and display web-based content going forward.
