Wikipedia commemorates 25th anniversary by inking AI licensing deals
Wikipedia, the ninth most visited site globally, boasts over 65 million articles in 300 languages, maintained by some 250,000 volunteers
Wikipedia has marked its 25th anniversary by announcing a series of new business deals with several prominent artificial intelligence companies, including Amazon, Meta Platforms, Perplexity, Microsoft, and France's Mistral AI.
The online crowdsourced encyclopedia, a long-standing pillar of the early internet, is now navigating a digital landscape increasingly dominated by Big Tech and generative AI chatbots.
These new licensing agreements come as questions mount over who should bear the cost of the AI boom, particularly given the aggressive data collection methods employed by AI developers, often drawing from Wikipedia's vast, freely available knowledge base. The Wikimedia Foundation, the non-profit organisation behind the site, previously secured Google as a customer in 2022 and later struck deals with smaller AI players like the search engine Ecosia.
The latest partnerships are set to help one of the world's most visited websites monetise the substantial traffic it receives from AI companies. These firms are now paying to access Wikipedia content "at a volume and speed designed specifically for their needs," according to the Wikimedia Foundation, which did not disclose financial specifics.
Despite legal battles elsewhere concerning copyright and AI training, Wikipedia founder Jimmy Wales expressed his approval. "I'm very happy personally that AI models are training on Wikipedia data because it’s human curated," Mr Wales told The Associated Press.
"I wouldn’t really want to use an AI that’s trained only on X, you know, like a very angry AI," he added, referencing billionaire Elon Musk's social media platform. Mr Wales emphasised the site's desire to collaborate with AI companies rather than obstruct them, but insisted they "should probably chip in and pay for your fair share of the cost that you’re putting on us."
Last year, the Wikimedia Foundation urged AI developers to pay for access via its enterprise platform, noting an 8% drop in human traffic. Concurrently, visits from bots, often disguised to avoid detection, were heavily straining its servers as they scraped content to feed large language models.
This trend underscores a shift in online behaviour, with AI overviews and chatbots summarising information instead of directing users to original sites.
Wikipedia, the ninth most visited site globally, boasts over 65 million articles in 300 languages, maintained by some 250,000 volunteers.
While its content is free for all, its underlying infrastructure is not. "But our infrastructure is not free, right?" Wikimedia Foundation CEO Maryana Iskander said in an interview from Johannesburg, South Africa. She explained that maintaining servers and other infrastructure, which enables both individuals and tech companies to "draw data from Wikipedia," incurs significant costs. Ms Iskander is due to step down on 20 January, to be succeeded by Bernadette Meehan.
The majority of Wikipedia's funding comes from 8 million individual donors. Mr Wales highlighted their expectations: "They're not donating in order to subsidize these huge AI companies." He added, "You know what, actually you can’t just smash our website. You have to sort of come in the right way."
Looking ahead, the Wikimedia Foundation has outlined an AI strategy that Mr Wales believes could introduce tools to alleviate tedious tasks for editors. While AI is not yet capable of writing Wikipedia entries from scratch, it could, for instance, update dead links by scanning text and finding alternative sources. "We don’t have that yet but that’s the kind of thing that I think we will see in the future," Mr Wales predicted.
Artificial intelligence could also enhance the Wikipedia search experience, evolving from traditional keyword searches to a more conversational, chatbot-like interface. "You can imagine a world where you can ask the Wikipedia search box a question and it will quote to you from Wikipedia," Mr Wales explained.
It could respond by saying "here’s the answer to your question from this article and here’s the actual paragraph. That sounds really useful to me and so I think we’ll move in that direction as well."
Reflecting on Wikipedia's early days, Mr Wales described it as a thrilling period, with many motivated to contribute after he and co-founder Larry Sanger launched it as an experiment. However, he acknowledged that the early internet also had a darker side. "People were pretty toxic back then as well. We didn’t need algorithms to be mean to each other," he said. "But, you know, it was a time of great excitement and a real spirit of possibility."
More recently, Wikipedia has faced criticism from the political right, with some dubbing it "Wokepedia" and accusing it of left-leaning bias. Republican lawmakers in the US Congress are investigating alleged "manipulation efforts" in its editing process, claiming these could undermine neutrality and impact AI systems relying on its data.
Elon Musk, a notable critic, last year launched his own AI-powered rival, Grokipedia, and has accused Wikipedia of being filled with "propaganda," urging people to cease donations. However, Mr Wales dismissed Grokipedia as a "real threat," explaining that large language models, on which it is based, are not yet capable of producing high-quality reference material. "So a lot of it is just regurgitated Wikipedia," he stated. "It often is quite rambling and sort of talks nonsense. And I think the more obscure topic you look into, the worse it is." He clarified that his criticism was not specific to Grokipedia, but rather "just the way large language models work."
Mr Wales mentioned knowing Mr Musk for years but noted they haven't been in touch since Grokipedia's launch. "I should probably ping him," he mused, adding that his message would simply be: "'How’s your family?' I’m a nice person, I don’t really want to pick a fight with anybody."
Bookmark popover
Removed from bookmarks