Wikipedia just drew a clear line for AI builders: if your app or model relies on Wikipedia, use the official pipes, credit the people who wrote the articles, and stop scraping the site like a mystery bot.
In a Monday blog post, the Wikimedia Foundation asked developers to route heavy use through Wikimedia Enterprise, its optional paid service, and to give proper attribution to Wikipedia editors. The idea is simple. Use the Wikipedia API built for scale, say where the facts came from, and keep the servers healthy so volunteers and readers are not pushed aside.
What is Wikimedia Enterprise? It is a paid, opt‑in product that delivers clean, reliable, high‑volume access to Wikipedia content. Companies get fast feeds and bulk data without hammering the public site. The money helps fund the nonprofit’s mission. That feels fair, especially when your product is pulling millions of requests a day.
The Foundation also flagged a worrying pattern. After it improved bot detection, it found that unusual traffic spikes in May and June came from AI bots that tried to look human. At the same time, human page views fell 8% year over year. That drop matters. Wikipedia survives on people. Fewer visits can mean fewer edits, fewer new contributors, and fewer small donations that keep the lights on.
Attribution is not just a nice‑to‑have. Wikipedia’s content is under a Creative Commons Attribution‑ShareAlike license. If you reuse it, you must credit the source clearly, link back to the article, and respect the license terms. In AI products, that means telling users where the information comes from and giving them a way to click through and read more. It also means crediting the human contributors who did the hard work.
There is a path that fits different needs. For small tools and research projects, Wikipedia’s free APIs and data dumps remain available. For large products and model providers, the Wikimedia Enterprise API is the right choice. It is designed for scale and reliability, and it avoids placing heavy load on the public site. If you run a large model or a popular assistant, this is the route to take.
If you build AI and want to do this right, here is the short checklist:
- Use the Wikimedia Enterprise API for high‑volume or commercial use.
- Give clear credit in your UI, link to the specific Wikipedia article, and keep the license notice.
- Identify your bot with a truthful user agent, respect robots.txt, and stick to rate limits.
- Cache and mirror responsibly so you are not hitting Wikipedia on every single request.
One more note that made me smile a little. The Foundation is not anti‑AI. Earlier this year, it outlined an AI strategy for editors that focuses on practical help. Think translation, cleanup, and help with repetitive tasks. AI assists the humans, it does not replace them.
The bottom line: Wikipedia is asking for responsible use, not special treatment. If your business benefits from Wikipedia, use the proper channels, support the nonprofit, and give credit where it is due. Your users get clearer sourcing, volunteers get the recognition they deserve, and the encyclopedia stays strong for the long run. Honestly, that is a win for everyone.
Also Read | OpenAI Locks in $38 Billion AWS Deal to Run ChatGPT on Hundreds of Thousands of Nvidia GPUs
Source: Wikimedia Foundation blog
