Wikipedia Traffic Fell Amid Automated AI Scrapers

The platform recorded an 8% dip in human page views between 2025 and 2024 as AI tools bypassed traditional navigation.

Updated on Oct. 2, 2026 in Artificial Intelligence

Bold flat-color editorial illustration of a glass prism on a plinth refracting light, representing the fragmentation of web traffic by AI tools.
Wikipedia experienced an 8% decline in human traffic between 2024 and 2025 as AI scrapers bypassed traditional navigation to generate direct summaries. AI Illustration. Upload story photo >

Live Poll

Do you believe AI summaries are making it harder for independent websites to survive?

Human traffic to Wikipedia declined by 8% between March 2025 and August 2025 compared to the same period in 2024. This trend follows the deployment of AI scrapers that generate summaries directly from site content, often preventing users from visiting the actual pages.

Why it matters

The shift highlights how AI-driven search models are fundamentally decoupling content consumption from traditional web traffic patterns. This creates a reliance on intermediary AI systems that capture value without necessarily directing audiences to the original source.

The Wikimedia Foundation recorded an 8% year-over-year drop in human visitors while bandwidth consumption for media downloads grew by 50% since January 2024. These figures reflect a transition where automated crawlers now account for significant infrastructure load previously associated with human users.

The players

Wikimedia Foundation

A non-profit organization that manages the hosting, development, and server infrastructure for Wikipedia and related open-knowledge projects.

Bernadette Meehan

The current CEO of the Wikimedia Foundation who previously served as the US Ambassador to Chile from 2022 to 2025.

Google

A search engine and technology platform that historically accounted for over 90% of traffic directed to Wikipedia pages.

The details

The foundation identified the decline by implementing upgraded bot-detection filters, a software layer that distinguishes automated traffic from human browsing. AI tools now generate summaries using Wikipedia as a source, effectively bypassing the 90% of traffic that traditionally arrived via Google search. This process, known as scraping, occurs when automated bots extract and aggregate data without returning a referral click to the original domain.

Timeline

  1. January 2024: Bandwidth usage for media downloads began climbing.

  2. March 2024 – August 2024: The baseline period used for the traffic comparison.

  3. March 2025 – August 2025: The window in which human page views fell by 8%.

  4. October 17, 2025: The foundation publicly reported the traffic decline.

  5. January 20, 2026: Bernadette Meehan assumed the role of CEO at the Wikimedia Foundation.

The Tech Race

This trend mirrors the broader struggle between information creators and AI aggregation models that now prioritize direct summary delivery. The foundation faces the challenge of maintaining site visibility in a landscape where search engines increasingly favor internal AI responses over traditional outbound links.

Users will increasingly see Wikipedia-derived facts presented directly within AI chat interfaces rather than through traditional web searches. While this streamlines access to information, it diminishes the visibility of the primary site and alters how knowledge is attributed and referenced online.

The takeaway

The data confirms that automated scrapers are fundamentally reshaping how users access encyclopedic knowledge. Readers should monitor if the foundation introduces new licensing or technical access policies to manage the increased bandwidth costs associated with AI-driven traffic.

Further reading

For broader trends on how generative models are shifting internet navigation, explore our Artificial Intelligence coverage.

Source note: This article includes information reported by Crypto Briefing.

Live Poll

Do you believe AI summaries are making it harder for independent websites to survive?