Wikipedia's volunteer editor community voted 44-2 to prohibit all AI-generated text from the encyclopedia's 7.1 million English-language articles. The policy, which passed on March 20, 2026, establishes one of the internet's most visited platforms as a clear holdout against AI automation.
Text generated by large language models often violates several of Wikipedia's core content policies, the updated policy states. The previous guidance merely suggested LLMs should not be used to generate new Wikipedia articles from scratch. The new language is unambiguous: the use of LLMs to generate or rewrite article content is prohibited.
"Text generated by large language models (LLMs) often violates several of Wikipedia's core content policies."
Two narrow exceptions persist. Editors may use LLMs for basic copyediting of their own writing, and for translation assistance, provided they are fluent in both languages and verify accuracy.
The decision reflects practical experience. Editors documented repeated instances of AI systems fabricating citations, inventing biographical details, and creating plausible but entirely fictional information. Generating AI content takes seconds, but identifying errors takes hours — a disproportionate burden on unpaid volunteers.
The policy applies specifically to English Wikipedia. Some other language editions may decide differently; Spanish Wikipedia has fully banned LLMs. The vote positions Wikipedia as a deliberate holdout in an increasingly AI-integrated internet, prioritizing accuracy and community stewardship over efficiency.