The Wikimedia Foundation has partnered with several artificial intelligence companies, including Ecosia, Microsoft, Mistral AI, Perplexity, Pleias, and ProRata, to allow the use of Wikipedia content for training AI models. These agreements, facilitated through Wikimedia Enterprise, aim to enhance the foundation’s long-term sustainability as online behaviors shift toward AI-powered answers and fewer direct Wikipedia visits. Wikipedia remains one of the largest, most-used websites globally, offering over 65 million articles in more than 300 languages, but it has recently experienced an 8% decline in human visits, partly due to AI-generated summaries reducing direct traffic. These content licensing deals occur amid growing legal and ethical debates over how AI companies source and use data for model training. While some organizations, like Wikimedia and Reddit, have reached licensing agreements with tech firms, others—including book publishers, writers, and entertainment companies—are pursuing lawsuits alleging copyright infringement. Recent legal outcomes in the U.S. have delivered mixed results, with courts sometimes ruling in favor of fair use by AI companies, but criticism continues over the unauthorized use of protected works. Broader regulatory scrutiny and industry calls for clearer standards on AI training practices are mounting worldwide.