Wikipedia Now Bans AI-Generated Articles

First Bandcamp, and now Wikipedia… The Verge reports that last week, the online encyclopedia adopted a new policy concerning articles written with AI large language models (LLMs). Noting that AI-generated content can violate the site’s core content policies (e.g., “all material in Wikipedia must be attributable to a reliable, published source”), Wikipedia now officially prohibits “the use of LLMs to generate or rewrite article content.”

There are two important exceptions, however. First, editors can use LLMs to suggest and incorporate “basic copyedits” into their own writing, but those edits must be reviewed by a human to ensure that “the LLM does not introduce content of its own.” And second, LLMs can be used to translate non-English Wikipedia articles into English, but those changes must adhere to Wikipedia’s LLM-assisted translation policies (e.g., the editor must be “skilled enough in both the origin language and English to confirm the translation is accurate”).

It really has been fascinating to watch the evolution of Wikipedia over the years. I can still remember a time not too long ago when Wikipedia was viewed with no small amount of skepticism. Sure, you might use it for something trivial, but you’d never cite it as a reliable source. But to the credit of Wikipedia’s army of editors, they’ve stayed the course and stuck to their (sometimes byzantine) principles. Even with its flaws, the site now often feels like one of the last remaining bastions of truth in a world that’s increasingly awash in AI slop and hallucinations.