Wikimedia Expands Collaborations With Amazon, Meta, Microsoft On AI Data Access

In Brief

Wikipedia marks 25 years by partnering with major AI companies to provide human-curated data for model training, highlighting its role in powering AI.

Nonprofit organization behind the Wikipedia free-content encyclopedia, Wikimedia Foundation, announced that the project has turned 25, marking a quarter-century of providing reliable, human-curated knowledge to users worldwide. In conjunction with its 25th anniversary, Wikipedia unveiled partnerships with several AI companies, including Amazon, Meta, Microsoft, Mistral AI, and Perplexity.

Through Wikimedia Enterprise, these companies will pay to access Wikipedia’s data for the development and training of AI models, offering an alternative to web scraping.

Wikimedia noted that these partnerships were formalized over the past year but had not been publicly disclosed until now. According to the organization, all participating companies use Wikimedia Enterprise to integrate human-governed knowledge into their platforms at scale.

These tech firms join existing partners such as Ecosia, Pleias, ProRata, and Google, the latter being one of Wikimedia Enterprise’s first collaborators in 2022.

Wikipedia’s content powers generative AI chatbots, search engines, voice assistants, and other technologies. The long-term development of AI and technology platforms relies on sustaining projects like Wikipedia, which provide the human knowledge foundational to these systems.

AI Boom Raises Questions Over Use Of Human-Generated Content

The recent surge in AI has brought data rights to the forefront, sparking legal and ethical questions about the use of human-generated content from platforms such as Wikipedia and Reddit.
Last year, Elon Musk introduced Grokipedia, an AI-powered alternative to Wikipedia. Marketed by him as less biased and “anti-woke,” Grokipedia features entries generated entirely by xAI’s large language model, Grok.