The Wikimedia Foundation, the nonprofit organization behind Wikipedia, has entered the copyright dispute between The New York Times, OpenAI and Microsoft. It has urged a US court to adopt a balanced approach when deciding whether artificial intelligence companies can use copyrighted material to train their models and generate responses.
Wikimedia filed an amicus brief, also known as a friend-of-the-court submission, in the ongoing copyright litigation in New York. The organization believes that an overly restrictive or overly permissive ruling could damage the open internet and limit access to reliable information.
Wikimedia is not fully supporting either the publishers or the AI companies. Instead, it wants the court to distinguish between training AI models and using online content to generate chatbot responses.
The foundation argues that using copyrighted material to train large language models (LLMs) should generally qualify as fair use when the resulting AI system creates something new and different from its original sources.
However, it takes a different position on retrieval-augmented generation (RAG). This technology allows chatbots to retrieve information from websites and use it to answer user questions.
Wikimedia argued that when AI systems reproduce existing articles through RAG, companies should respect the original publishers' licensing conditions rather than automatically claim fair use.
Also Read: OpenAI and Copyright: The Battle Between AI and Creative Rights
Wikipedia relies on thousands of volunteer contributors who create and maintain its articles. The platform has more than 67 million articles across 300 languages. Its content is available under the Creative Commons CC BY-SA 4.0 licence. This allows reuse and adaptation, but requires proper credit to original contributors and similar licensing for adapted material.
Wikimedia believes AI companies should follow these conditions when retrieving and reproducing Wikipedia content. Without attribution, chatbots could use information without recognising the people who created it.
Also Read: No Copyright Breach, Delhi High Court Backs OpenAI Training
The original lawsuit was filed by The New York Times in 2023 against OpenAI and Microsoft, alleging copyright infringement through the use of published material. Both companies have argued that using publicly available content for AI development falls under fair use.
Wikimedia warns that allowing unrestricted copying could weaken publishers and other information sources. On the other hand, restricting AI training too heavily could slow innovation and limit the development of open-source technologies. The court's decision could influence how AI companies collect, process and present online information in the coming years.