Wikipedia co-founder, Jimmy Wales, has said the online encyclopedia will not allow artificial intelligence (AI) systems to directly edit articles on its platform due to concerns over the reliability of AI-generated information.
Wales made the remarks while speaking to journalists on the sidelines of a climate action week event in London.
According to him, although recent advances in AI technology have reduced the frequency of so-called “hallucinations”—instances where AI systems generate false information and present it as factual—the problem remains a significant challenge.
“The problem of AI hallucinations is still very, very bad,” Wales said.
He explained that Wikipedia’s commitment to accuracy and verifiability means AI systems cannot yet be trusted to independently make changes to articles.
“We would not let it edit directly because you can’t really trust it enough,” he added.
However, Wales noted that AI could still be useful in supporting Wikipedia’s network of volunteer editors by helping them identify specialised news stories, research developments and other topics that might otherwise escape attention.
He also highlighted the growing dependence of artificial intelligence companies on Wikipedia’s content, noting that many AI-powered platforms rely on information from the encyclopedia to answer user queries.
According to Wales, traffic from AI bots accessing Wikipedia has increased substantially, while visits from human users have declined by approximately eight per cent.
He described the trend as significant but not alarming, noting that Wikipedia remains among the world’s ten most-visited websites.
Wales said the platform’s reliance on user donations rather than advertising revenue makes it less vulnerable to fluctuations in web traffic.
He further called on AI companies to contribute financially to support the infrastructure and resources they utilise.
“We encourage them to pay their fair share because hammering us with millions of requests costs real money,” he said.
The Wikipedia co-founder disclosed that the Wikimedia Foundation had already entered into agreements with several major technology firms regarding the use of Wikipedia’s content and server resources.
He added that the foundation was taking steps to address unauthorised or excessive use of its systems.
“We’re starting to block the ones who aren’t behaving themselves, but we’ll see how that goes,” Wales said.
Industry experts say the comments reflect growing debates over the relationship between AI developers and content providers, particularly concerning accuracy, copyright, infrastructure costs and the sustainability of freely accessible online knowledge resources.
