Wikipedia co-founder Lawrence Mark Sanger has warned that the online encyclopedia has become a propaganda tool, saying the Central Intelligence Agency (CIA) and corporate edits have undermined its neutrality.
“Wikipedia was recognized early on as an important propaganda organ,” Sanger said during an interview with German newspaper Berliner Zeitung on Friday.
“What we know as fact, based on all sorts of evidence, is that the CIA has already edited Wikipedia,” Sanger added.
Evidence of CIA and Federal Bureau of Investigation (FBI) edits emerged in 2007 after programmer Virgil Griffith created WikiScanner, a tool that linked Wikipedia edits to the Internet Protocol (IP) addresses of organizations.
He further referred to reporting by the Bureau of Investigative Journalism (BIJ), which found that communications company Portland Communications had rewritten Wikipedia entries for governments and wealthy institutions, including Qatar, the Bill & Melinda Gates Foundation and Libya’s Sovereign Wealth Fund.
Sanger also pointed to former Wikimedia Foundation chief Katherine Maher, who told the Atlantic Council in 2021 that the organization had held discussions with governments and other online platforms.
“The powerful are very interested in Wikipedia,” Sanger noted, arguing that institutions such as the World Economic Forum and World Bank were increasingly treated by the platform as authoritative sources.
Wikipedia’s influence now extends beyond the website because search engines and AI chatbots increasingly use its content for training, he warned.
“Such models do not develop a life of their own once they are launched,” Sanger explained, arguing that AI systems remain heavily dependent on their training data.
Sanger also noted that Wikipedia’s coverage of the Israeli genocide in Gaza has been “clearly very one-sided.”
Israel has spent more than $1 billion on international propaganda efforts since October 7, 2023, including campaigns designed to influence information available to AI systems.
Part of that money goes to what experts call “LLM poisoning,” which involves building pro-Israel websites designed to feed chatbots such as Claude, ChatGPT, and Grok with Israeli-approved information when users ask about the genocide.