Wikipedia bans AI-generated articles, AI may only assist

English-language Wikipedia instructs its authors not to use AI models for article creation. This is because texts created using artificial intelligence (AI) often violate several Wikipedia content guidelines. However, there are exceptions where AI can be used, such as for translations and general text editing. Both authors and moderators are subject to specific rules in these cases.

The German-language Wikipedia has already decided on a comprehensive AI ban after a majority of almost two-thirds of participants adopted the proposal for new AI guidelines in February. Accordingly, articles generated with large language models (LLMs) are expressly forbidden. Exceptions here also include machine translations that rely on AI, AI-assisted error detection, and pure research with AI.

AI ban now also in English

English-language Wikipedia has now also integrated this into its guidelines. This prohibits articles that have been completely generated by AI or rewritten with the help of AI. However, Wikipedia warns that detecting purely AI texts can be difficult. This is because some people “may have similar writing styles to LLMs.” Therefore, articles should not be rejected solely because of an AI-like style; editors should focus on the “text's compliance with core content policies” and consider the last changes made by the author in question before rejecting an article due to the AI ban.

AI is not fundamentally demonized either. Language models are allowed, for example, for post-editing texts, as long as the AI “does not introduce content of its own.” AI-assisted machine translations of Wikipedia articles from other languages are also accepted. However, it is assumed that the editor masters the foreign language sufficiently to confirm that the translation is correct. In addition, any AI hallucinations in the text should be removed and the sources of the original article checked.

AI Flood: First Speedy Deletion, Now Prevention

The new AI guidelines are a further reaction to the increase in AI articles on the online encyclopedia. Back in August 2025, Wikipedia reacted to the AI flood with an adjusted guideline for speedy deletion. This allows administrators to delete obviously machine-generated and unchecked contributions without the usual one-week discussion period if they meet certain criteria.

Normally, deleting a Wikipedia article requires a consensus found in a seven-day discussion. If articles show obvious signs of an AI model, they can be deleted immediately, as was already the case for clear cases like vandalism or pure advertising. With the guideline now extended to include the prohibition of AI-generated articles, Wikipedia aims to act preventively in both English-speaking and German-speaking regions, rather than just being able to delete faster.

(fds)