Video: AI is CRASHING Wikipedia! 🤯 50% Bandwidth Spike! - YouTube
AI bots have significantly impacted Wikipedia's bandwidth, leading to a 50% surge since January 2024. This surge is primarily due to automated bots scraping data for AI model training, which has strained Wikimedia's servers and increased bandwidth usage for downloading multimedia content.
Wikimedia Foundation, which hosts Wikipedia and Wikimedia Commons, has reported that these bots are vacuuming up terabytes of data through direct crawling, APIs, and bulk downloads. This non-human traffic has imposed steep technical and financial costs on the foundation, often without proper attribution that supports Wikimedia's volunteer ecosystem.
One notable incident occurred when former US President Jimmy Carter passed away in December 2024. While his Wikipedia page drew millions of views, the simultaneous streaming of a 1.5-hour video from Wikimedia Commons doubled the normal network traffic, temporarily maxing out several Internet connections. This event highlighted the underlying issue of baseline bandwidth being consumed by bots scraping media at scale.
To mitigate these challenges, Wikimedia engineers have been forced to reroute traffic and implement various technical solutions. However, the foundation continues to face difficulties as many AI-focused crawlers ignore robots.txt directives, spoof browser user agents, and rotate through residential IP addresses to evade detection.
Wikimedia is now focusing on systemic approaches under the initiative WE5: Responsible Use of Infrastructure, aiming to establish sustainable boundaries while preserving openness. The foundation emphasizes that while its content is freely licensed, the infrastructure is not, and better coordination between AI developers and resource providers is essential to resolve these issues.
For more detailed information, you can visit the TechTimes article or the Ars Technica article.