You know that moment when you’re researching something important and stumble across an article that reads like it was written by a committee of robots having a stroke? Wikipedia’s volunteer editors are officially done with that nonsense.
On March 20, 2026, English Wikipedia’s editorial community voted 44-2 to ban large language models from generating or rewriting article content. The decision marks a decisive shift from earlier guidelines that only prevented AI from creating entirely new articles.
The Vote That Wasn’t Even Close
English Wikipedia editors approved sweeping AI content restrictions with overwhelming support.
The new policy covers the whole enchilada—no ChatGPT drafts, no Claude rewrites, no algorithmic shortcuts to encyclopedia entries. The exceptions remain surgical: editors can use LLMs for basic copyediting suggestions on their own writing, but only after human review.
Translation assistance gets similar narrow approval, though the policy warns that AI “can go beyond what you ask… changing the meaning” unsupported by sources. The restrictions reflect growing concerns about content integrity as AI tools become more sophisticated.
When Citation Hallucinations Meet Reality
Enforcement targets specific behaviors that compromise Wikipedia’s reliability standards.
WikiProject AI Cleanup—yes, that’s a real volunteer group—has been handling an increasing flood of AI-generated errors.
These range from phantom citations to mass-produced stub articles that sound authoritative but lack substance. Administrators can now block or topic-ban users based on output quality rather than just relying on detection tools. Smart move, considering AI detection remains about as reliable as predicting TikTok trends.
The Domino Effect Begins
Wikipedia’s decision could influence how other platforms handle AI-generated content floods.
This isn’t happening in a vacuum. StackOverflow already implemented similar restrictions after their platform got swamped with AI-generated coding answers that looked helpful but often contained subtle errors.
Lebleu predicts a “domino effect” empowering online communities to push back against AI content floods. The precedent matters because Wikipedia serves as training data for the very AI systems now banned from contributing to it.
Your daily information diet just got more reliable, assuming Wikipedia’s volunteer enforcers can keep up with increasingly sophisticated AI output. The encyclopedia that anyone can edit just reminded everyone that “anyone” still means humans.
