Wikipedia, long regarded as one of the internet’s few remaining public-interest platforms, is now seeing a notable dip in human readership. According to a new update from the Wikimedia Foundation, the site’s human pageviews dropped by 8% year over year—a decline that appears tied to the changing ways people consume and access information online.
Marshall Miller, a Wikimedia Foundation executive, explained that the decrease became apparent after a recent update to the site’s bot detection systems. The adjustment revealed that what had previously looked like a surge in traffic during May and June was largely the result of automated systems designed to evade detection. Once that traffic was properly filtered, the numbers showed a clear downward trend in genuine human visits.
Miller attributes the decline to two overlapping shifts: the rise of generative AI search results and the growing dominance of social video platforms as primary information sources. Search engines are increasingly offering AI-generated summaries that answer users’ questions directly, reducing the need to click through to original sources like Wikipedia. At the same time, younger users are relying more heavily on platforms such as TikTok, YouTube Shorts, and Instagram Reels for quick explanations and cultural context—spaces where Wikipedia has little visibility. Google has pushed back on claims that its AI-generated search answers are diminishing referral traffic, but Wikipedia’s numbers suggest a real impact nonetheless.
Despite this, Miller maintains that Wikipedia’s underlying influence remains significant. Much of the data and factual content used by generative AI systems originates from the encyclopedia’s vast database. In other words, Wikipedia’s work continues to inform global knowledge even when users don’t visit the site directly. The organization has experimented with its own AI-generated summaries in the past but paused the project after community members raised concerns about accuracy and editorial oversight.
Still, the shift poses deeper risks for the open web. Wikipedia relies heavily on its community of volunteer editors and small donations from individual readers. A drop in direct engagement could lead to fewer contributors, slower content updates, and less financial support. Miller warns that if users grow less aware of where their information originates, the ecosystem that sustains reliable, human-curated knowledge may weaken.
To address this, Wikipedia is developing new frameworks for content attribution and launching initiatives aimed at expanding its reader base. Two internal teams are working to make the site more accessible and relevant to new audiences, and the foundation is actively seeking volunteers to join these efforts. Miller also called on tech companies that use Wikipedia content to drive more traffic back to the source, ensuring that the encyclopedia’s editors and infrastructure remain supported.
In closing, Miller urged readers to engage critically with information online: to follow citations, read original sources, and recognize the human labor behind the data that fuels AI systems. His message underscores a growing tension across the digital landscape—between convenience-driven automation and the collective human effort that still underpins the web’s most trusted knowledge repositories.
