Wikimedia Foundation Intervenes in Landmark AI Copyright Lawsuits
The organization advocates for a nuanced fair use standard to protect the open web.
1Wikimedia's Strategic Legal Intervention
The Wikimedia Foundation has filed an amicus brief in the ongoing copyright litigation between the New York Times, OpenAI, and Microsoft. The nonprofit seeks a middle ground, arguing that current legal arguments from both sides are overly simplified. They aim to protect the long-term sustainability of free, open-source knowledge platforms.
2Defining Fair Use Boundaries
Wikimedia argues that training large language models on copyrighted data should qualify as fair use. However, they draw a firm line when AI systems regurgitate that content to users. They contend that direct reproduction of material without adhering to original licensing terms undermines the collaborative communities that create the data.
3The Risk of Overbroad Rulings
The organization warns that a black-and-white court ruling could have existential consequences for the open web. If courts restrict all AI training, it could deprive Wikipedians of essential tools. Conversely, allowing unrestricted regurgitation could discourage human authors from contributing, ultimately fragmenting the flow of knowledge between human creators and their audiences.
4AI as a Tool for Editors
Despite copyright concerns, Wikimedia acknowledges that large language models provide significant utility for Wikipedia operations. These systems currently assist human editors in detecting vandalism, identifying potential doxxing, and flagging articles that require improvement. Balancing these technological benefits against copyright protection remains a primary challenge for the foundation's future policy strategy.
5The New York Times Position
The New York Times maintains that OpenAI infringed on its copyright by scraping website content to build training datasets. They argue against an expansive definition of fair use for LLMs. This litigation highlights the tension between publishers seeking to protect their intellectual property and tech companies pursuing massive data-driven model development.
6Microsoft and OpenAI Defense
OpenAI and Microsoft contend that their LLM training constitutes fair use. OpenAI claims that regurgitating full content is merely a rare bug rather than a core feature. Microsoft cited expert analysis suggesting that Copilot rarely reproduces full sentences, arguing that this behavior does not undermine the transformative purpose of their training models.
7Proposed Licensing Compromise
Wikimedia proposes that if AI models output material that supplants reader engagement with original sources, those companies must respect licensing requirements. For Wikipedia, this means enforcing Creative Commons attribution. They argue that AI output must transform the original work significantly to qualify as fair use rather than acting as a direct substitute.
8Preserving the Information Ecosystem
CEO Bernadette Meehan emphasizes that the foundation's position aims to defend the human knowledge Wikipedia relies upon. Finding a balance is a shared responsibility for the entire information ecosystem. The foundation believes that respecting contributors is essential to ensure that AI development does not stifle the very sources it depends on.
Original article: The Verge — Wikimedia Foundation Intervenes in Landmark AI Copyright Lawsuits