Microsoft and Meta join Wikipedia’s growing AI data ecosystem

Wikipedia is taking a new approach to make money by charging big tech firms for the massive amount of data they use. On Thursday, the group that runs the site announced a string of deals with names like Microsoft, Meta, and Amazon.

For years, the Wikimedia Foundation has watched tech giants use its free articles to build their products. Now, the non-profit is finally turning that reliance into a source of income. Over the last year, they have signed up newer players like the AI startup Perplexity and France’s Mistral AI. These companies join a growing list that already includes Meta and Amazon, as well as Google, which signed a deal back in 2022.

High costs of AI training

Wikipedia’s data is a tech treasure trove. Chatbots and virtual assistants are trained to speak and respond to inquiries using its 65 million articles, which are authored in more than 300 languages. However, there is an additional expense. Wikipedia’s servers are heavily taxed when these businesses “scrape” or extract massive volumes of data from the website. These expenses are typically covered by modest contributions from the general public, but the increased demand from AI businesses has increased those expenses.

To solve this, Wikimedia is pushing its “Enterprise” service. It allows big companies to pay for the content they need in a format that works better for their high-tech systems. Lane Becker, who leads the Enterprise branch, said these companies realize they need to help fund the site if they want it to survive. He said it took some time to figure out exactly what features to offer to get companies to switch from the free site to the paid version.

Supporting the volunteer community

The site itself is still kept running by a massive team of 250,000 volunteers who write and check the facts for free. Tim Frank from Microsoft said that having access to honest, high-quality info is a big part of how they see the future of AI. He noted that by working with Wikimedia, they are helping to keep a system where the people who write the content are still supported.

Amidst these business changes, the foundation is also getting a new leader. Bernadette Meehan, a former U.S. Ambassador, is set to take over as chief executive on January 20.