An institution of the open internet is showing signs of strain as the very technologies it helped to build begin to alter the digital landscape. The Wikimedia Foundation, the non-profit organization behind Wikipedia, has reported a significant downturn in traffic, with human pageviews dropping by approximately eight percent compared to the previous year.[1][2][3][4][5][6][7] This decline, which became apparent after the foundation updated its systems to better identify and exclude sophisticated bot traffic, points to a fundamental shift in how people access information online. The primary drivers of this change are the rise of generative artificial intelligence and the enclosed ecosystems of social media platforms, both of which are increasingly providing direct answers to users, often without a click-through to the original source.

The drop in visitors was quantified after a revealing discovery. Around May 2025, the foundation noticed an unusual spike in what appeared to be human traffic, largely originating from Brazil.[1][5][8][9] An investigation led to an update in bot detection systems, which determined that much of this traffic came from bots designed to mimic human behavior to evade detection.[2][5][10][8] After reclassifying data from March to August 2025, the underlying trend became clear: genuine human visits were down.[2][5][11] This decline is not happening in a vacuum; many publishers and content platforms are reporting similar shifts as users spend more time getting information directly from search engines, AI chatbots, and social platforms.[5][6][11] Younger users, in particular, are increasingly turning to social video platforms like TikTok, YouTube, and Instagram for knowledge, bypassing the open web altogether.[1][11][10][12][13]

At the heart of this issue is a complex and increasingly fraught relationship between Wikipedia and the AI industry. Large language models, the technology underpinning popular AI chatbots like ChatGPT and search summaries like Google's AI Overviews, have been trained extensively on Wikipedia’s vast, freely accessible repository of human-curated knowledge.[11][14][15] It is often the single most important source for training these models.[14] Now, these same AI systems are summarizing Wikipedia's content and presenting it directly to users, obviating the need for a visit to the encyclopedia itself.[2][3][6][7] This creates a parasitic dynamic where AI companies leverage Wikipedia's data for their commercial products while simultaneously eroding the platform's direct user engagement.[16][17] The Wikimedia Foundation has called on these tech giants to be more transparent about their sources and to provide clear opportunities for users to visit and participate in the platforms that provide the information.[3][11]

The implications of this traffic decline extend far beyond simple website metrics, posing a potential existential threat to Wikipedia's operational model. The platform relies on a virtuous cycle: visitors become editors, and a portion of users become donors who financially support the site's infrastructure and independence.[1][3][11] A sustained drop in traffic could break this cycle, leading to fewer new volunteer contributors to grow and maintain the encyclopedia's content and a decrease in the individual donations necessary to keep it running.[3][5][11][10] This threatens the very standards of verifiability, neutrality, and transparency that have made Wikipedia a trusted source of information powering countless other services across the internet.[1][3] The feedback loop is perilous: if traffic declines, leading to less content moderation and fewer updates, the quality of Wikipedia's data could degrade, which would in turn lower the quality of the AI models that train on it.[18][15]

In response to this changing internet, the Wikimedia Foundation is not standing still. Beyond calling for better practices from AI companies, it is actively exploring new strategies to engage audiences where they are.[1] This includes experimenting with projects aimed at younger audiences on platforms like YouTube and TikTok and developing new tools to improve the experience for readers and first-time editors.[1][19] The organization is also developing a new content attribution framework to better credit its content when it is used by third parties.[10][13] While the future remains uncertain, it is clear that Wikipedia's role is evolving. It may be shifting from a primary destination for information to a foundational, often invisible, data layer for the AI-driven internet—a quiet powerhouse shaping the future of knowledge even as fewer users walk through its virtual doors.[19] The challenge now is to ensure this foundational role is sustainable, preserving the human-powered ecosystem that has made it one of the most valuable resources of the digital age.