Wikipedia has officially banned the use of artificial intelligence (AI) to generate or rewrite article content. The new policy, now in effect on English Wikipedia, marks one of the biggest editorial decisions the platform has made in recent years. According to The Verge, Wikipedia updated its guidelines late last week to say that articles written by AI violate “several of Wikipedia’s core content policies.”
What is currently prohibited and two exceptions
Under the new rules, Wikipedia editors are not allowed to use AI tools to generate article content or rewrite existing text. The concern isn’t accuracy, but the potential for AI to subtly distort meaning.The policy itself states that the LLM “may change the meaning of the text beyond what is required and not supported by the cited sources.” In simpler terms, AI tools may produce text that reads smoothly and sounds authoritative, but it can change what the article actually says.However, Wikipedia has identified two specific situations in which AI tools are allowed. The first one is writing assistance. Editors can use LLM to suggest improvements to their writing, such as grammar checks or light stylistic polish. The second exception is translation. In this case, editors can use AI to create a first draft translation of content from another language, but only if they’re proficient enough in both languages to catch any errors.
Contents of update policy
Text generated by large-scale language models (LLMs) often violates some of Wikipedia’s core content policies. For this reason, you are prohibited from using LLM to generate or rewrite article content, with the exceptions noted below.Editors are allowed to use LLM to suggest basic copy edits to their writing and incorporate some of it after human review, unless LLM introduces original content. Care must be taken as LLM may change the meaning of the text beyond the user’s request in a way that is no longer supported by the cited sources.If you use LLM to translate Wikipedia articles in another language to English Wikipedia, you should follow the guidance provided in Wikipedia:LLM-assisted translation.Some editors may have a similar writing style to LLM. Justifying a sanction requires evidence that goes beyond mere stylistic or linguistic cues, and it’s best to consider the text’s compliance with core content policies and recent edits by the editor in question.
