The Wikimedia Foundation announced Thursday it has signed partnerships with Microsoft, Meta, Amazon, Perplexity, and Mistral AI for access to Wikipedia content for AI training purposes. The announcement coincided with Wikipedia's 25th anniversary. The
Wikimedia Foundation operates the online encyclopedia, which relies primarily on small donations for funding. The partnerships allow tech companies to access Wikipedia's content through Wikimedia Enterprise, a commercial product designed for large-scale data needs. This replaces the previous practice of companies freely scraping Wikipedia's servers for training data. Wikipedia contains 65 million articles across more than 300 languages, making it a significant source of training data for AI language models and chatbots.
The content is created and maintained by approximately 250,000 volunteer editors globally. Wikimedia Foundation previously announced a partnership with Google in 2022. The organization also signed agreements last year with smaller AI companies including search engine Ecosia. AI scraping has increased server costs for the nonprofit organization. Last year, Wikimedia Foundation reported that human traffic to Wikipedia had declined 8 percent, while bot traffic from AI companies taxed its servers as they collected content for training large language models. Wikimedia Enterprise generated $8.3 million in revenue during fiscal year 2024-2025, representing a 148 percent increase from the previous year. Tim Frank, Microsoft's Corporate Vice President, stated the partnership helps create a sustainable content ecosystem where contributors are valued. Wikipedia founder Jimmy Wales told The Associated Press he welcomes AI training on Wikipedia data because it is human-curated. Wales said companies should contribute financially to cover the costs they impose on Wikipedia's infrastructure. Wikimedia Foundation named Bernadette Meehan, former U.S. Ambassador to Chile, as its new chief executive, effective January 20.
Comments