Wikipedia opens its content to AI companies through licensing deals

Wikipedia, one of the world's largest sources of knowledge, is opening its doors to artificial intelligence. Wikimedia Enterprise, the commercial arm of the Wikimedia Foundation, has signed a series of licensing agreements with tech giants such as Microsoft, Meta and Amazon, as well as with promising startups Perplexity AI and Mistral AI. The agreements give AI companies access to Wikipedia's vast data set to train and improve their models.

For a long time, Wikipedia remained a free, open resource available to everyone. But as large language models (LLMs) have grown in popularity, demand for high-quality training data has risen sharply. Wikimedia Enterprise saw an opportunity to monetize its content without violating the principles of openness and accessibility that underpin Wikipedia. Paid licenses give companies structured, optimized access to the data, plus technical support, which considerably simplifies the process of training AI models.

What does this mean for the AI industry? First, it provides access to an enormous body of verified, structured information, which is critical for training high-quality, reliable LLMs. Wikipedia contains millions of articles in many languages, covering a wide range of topics, from history and science to culture and technology. Using this data should help AI models understand the world better and generate more relevant and accurate answers. Second, the licensing agreements provide a sustainable source of funding for the Wikimedia Foundation, which should allow it to keep supporting and developing Wikipedia as a global knowledge resource.

The move also carries potential risks. It is important that the use of Wikipedia content respects the principles of neutrality and objectivity at the heart of the encyclopedia. Situations in which AI models trained on Wikipedia data spread disinformation or biased opinions must be prevented. The Wikimedia Foundation needs to monitor the use of its content closely and respond promptly to any violations. It is also important to be transparent about which Wikipedia data is used to train which AI models.

Signing licensing agreements with AI companies is an important step for Wikipedia and for the whole artificial intelligence industry. It opens new opportunities for AI development, but demands a responsible approach and adherence to the principles of openness and neutrality. In the future, we will likely see other major data sources follow Wikipedia's example and start monetizing their content for AI training. This could lead to the emergence of a new market for AI data, which in turn would have a significant impact on the development of artificial intelligence technologies.

*Meta is designated an extremist organization and banned in Russia.