The Wikimedia Foundation has issued a new call to action: it’s asking AI companies to stop scraping Wikipedia content and instead access it via its commercial offering, Wikimedia Enterprise. This move signals a significant shift in how open-knowledge platforms expect large-scale data users to behave.
What is being asked and why
- Wikipedia says that many AI developers and companies are harvesting content by scraping its pages—often at large scale—and this practise is putting pressure on its infrastructure and community.
- The foundation notes that its human page-views have declined by around 8% year-on-year—a trend it links partly to AI tools answering user queries, reducing visits.
- It emphasises that the Wikimedia Enterprise API, though paid, allows for high-volume, reliable access while also supporting its non-profit mission and preserving the volunteer-editor ecosystem.
- Wikipedia asks for proper attribution whenever its content is used in AI systems or outputs, emphasising transparency and respect for the thousands of human contributors.
Five Key Implications
1. Financial & sustainability pressures
As part of its argument, the foundation says that fewer human visits and more automated scraping reduce the donations and volunteer edits that sustain Wikipedia’s ecosystem.
2. Infrastructure strain due to scraping
Large-scale automated scraping drives traffic and load on Wikipedia’s servers. The foundation detected spikes in bot traffic disguised as human, particularly in mid-2025.
3. Shift in access model for large-scale users
By recommending the paid Enterprise API over free scraping, Wikipedia is nudging a change in how commercial users obtain large datasets—moving from “free for all” to “structured, paid access for scale”.
4. Attribution and value recognition
Wikipedia emphasises that, even if content is licensed under open terms, large-scale commercial usage should still acknowledge the volunteer editors and infrastructure behind it.
5. Precedent for other open-knowledge platforms
This move may influence other platforms with large open-data sets to similarly demand structured access, attribution or payment from large scale AI uses—shifting norms in the AI training-data ecosystem.
What this means for AI companies & developers
- If you’re building models or services that rely on Wikipedia or similar open-community content, consider using Wikimedia Enterprise to avoid strain or future access limitations.
- Evaluate your data-access strategy: large scale scraping not only risks infrastructure issues but also may affect reputation and compliance with platform expectations.
- Ensure proper attribution of sources in outputs and, where required, verify you have the rights or usage terms for the content.
- Monitor potential changes in access terms: Wikipedia’s push may mark early policy shifts around data access, licensing or pricing for large-volume users.
- For startups / smaller developers: consider whether alternative data sources or partnering with Wikipedia’s ecosystem might offer better alignment with sustainability and attribution.
Things to watch & limitations
- Wikipedia is not explicitly threatening legal action at this point; the tone is more advisory and normative.
- This request currently focuses on large-scale, commercial usage; everyday visits, casual uses and small-scale scraping may not be heavily affected for now.
- For India and other global regions: while Wikipedia is global, usage terms, enterprise pricing, and regional data-access implications may differ; local developers need to check terms & regional offerings.
- The core licence behind Wikipedia remains open (CC BY-SA), so this is not a sudden “paywall” for regular users—but rather a shift in how large-scale commercial consumption is handled. Medium
Conclusion
Wikipedia’s call to “stop scraping, please pay or access properly” marks a noteworthy moment in the evolving relationship between open-data platforms and the commercial AI ecosystem. By pushing for paid API access and proper attribution, the Wikimedia Foundation is trying to protect its infrastructure, volunteer base and mission in an era of massive AI-driven demand. For AI builders, this is a reminder that open data doesn’t always mean unfettered free access—especially at scale.
