History
Beginnings
The use of AI to generate articles on Wikipedia started with the rise in popularity of chatbots like ChatGPT. One year later, in 2023, the Wikipedia community noticed the problem and created a special WikiProject named AI Cleanup to clean Wikipedia from AI content. The project wrote its own guidelines to help users in spotting AI content. As of 2025, it had created a list of over 500 articles pending review for suspected AI writing. Wikipedia has also created a special template for suspected AI-generated articles, which was used in articles like Danish nationalism and Natalie Portman. On October 2024, a study by Princeton University revealed that about 5% of 3,000 newly created articles (created on August 2024) on English Wikipedia were created using AI. The study said that some of the AI articles were on innocuous topics and that AI had likely only been used to assist in writing. For some other articles, AI had been used to promote businesses or political interests. [1] [2]
On December 6, 2022, a Wikipedia contributor named Pharos created the article "Artwork title" in his sandbox, declaring he used ChatGPT to experiment with it and would extensively modify it. He noted that the text needed to be toned down for neutrality. Another editor tagged the article as "original research", arguing that the article was initially unsourced AI-generated content, and sourced afterwards, instead of being based on reliable sources from the outset. Another editor who experimented with this early version of ChatGPT said that ChatGPT's overview of the topic was decent, but that the citations were fabricated. [3] Wiki Education Foundation reported that some experienced editors found AI to be useful in starting drafts or creating new articles. It said that ChatGPT “knows” what Wikipedia articles look like and can easily generate one that is written in the style of Wikipedia. It warned editors that ChatGPT had a tendency to use promotional language. [4] Miguel García, Wikipedia editor from Spain, said that when ChatGPT was originally launched, the number of AI-generated articles on the site peaked. He added that the rate of AI articles has now stabilized due to the community's efforts to combat it. He said that majority of the articles that have no sources are deleted instantly or are nominated for deletion. [5]
Signs of AI use and speedy deletion
On August 2025, the Wikipedia community created a policy that allowed users to nominate suspected AI-generated articles for speedy deletion. Editors usually recognize AI-generated articles because they use citations that are not related to the subject of the article or fabricated citations. The wording of articles is also used to recognize AI writings. For example, if an article uses language that reads like an LLM response to a user, such as "Here is your Wikipedia article on” or “Up to my last training update”, the article is typically tagged for speedy deletion. [1] [6] Other signs of AI use include: excessive use of em dashes, overuse of the word "moreover", promotional material in articles that describes something as "breathtaking” and formatting issues like using curly quotation marks instead of straight versions. During the discussion on implementing the speedy deletion policy, one user, who is an article reviewer, said that he is “flooded non-stop with horrendous drafts” created using AI. Other users said that AI articles have a large amount of “lies and fake references” and that it takes a significant amount of time to fix the issues. [7] [8]
Ilyas Lebleu, founder of WikiProject AI Cleanup, said that he and his fellow editors noticed a pattern of unnatural writing that they managed to connect to ChatGPT. He added that AI is able to mass-produce content that sounds real while being completely fake, leading to the creation of hoax articles on Wikipedia that he was tasked to delete. [9] [10] Wikipedia created a guide on how to spot signs of AI-generated writing. The guide states that AI uses editorial commentary in its content, it listed phrases like "it's important to note", "it is worth", and "no discussion would be complete without" as examples. It also said that AI uses phrases like "In summary", "In conclusion", and "Overall" in the end of the articles. The guide also said that fabricated sources are also a major sign of AI use, AI is known to create hallucinated sources that have fake DOIs or ISBNs, or broken 404 links. The guide also added that ChatGPT is known to add broken code while adding external links to articles, leaving "turn0search0" in its links. [11] The guide is called "List of signs of AI writing". [12]
Hoaxes and malicious AI use
In 2023, researchers discovered that ChatGPT can unintentionally fabricate information and make up fake articles for its users. At that time, ban on AI was deemed "too harsh" by the community. [13] [14] AI was deliberately used to create various hoax articles on Wikipedia. For example, Ilyas Lebleu and his team managed to expose a well-written 2,000-word article about an Ottoman fortress that never existed. The content of the article, while completely wrong, was difficult to debunk without knowledge of 13th-century Ottoman architecture. [15] [16] Another examle showed an user adding AI-generated misinformation to Estola albosignata, species of beetle. The paragraph the user cited and the citation looked normal, however, the source was not related to the subject at all and was about unrelated species of crab in French. [17]
AI has been used on Wikipedia to advocate for certain political viewpoints in articles covered by contentious topic guidelines. One instance showed a banned editor using AI to engage in edit wars and manipulate Albanian history-related articles. [18] Other instances included users generating articles about political movements or weapons, but dedicating the majority of the content to a different subject, such as by covering JD Vance or Volodymyr Zelensky in a non-neutral way. [19] Ilyas Lebleu said that there are many reasons to why some users add AI-generated content to Wikipedia, he said that they include deliberate vandalism with the intention to create a hoax, self-promotion or them falsely thinking that AI-generated content is correct. [20]
Reactions
In November 2023, Wikipedia co-founder Jimmy Wales said that AI is not a reliable source and that he is not going to use ChatGPT to write Wikipedia articles. [21] In July 2025, he proposed the use of LLMs to provide customized default feedback when drafts are rejected. [22]
Wikimedia Foundation product director Marshall Miller said that WikiProject AI Cleanup keeps the site's content neutral and reliable and that AI enables the creation of low-quality content. [23] When interviewed by 404 Media, Ilyas Lebleu described speedy deletion as a "band-aid" for more serious instances of AI use, and said that the bigger problem of AI use will continue. He also said that some AI articles are discussed for one week before being deleted. [24]
This page is based on this
Wikipedia article Text is available under the
CC BY-SA 4.0 license; additional terms may apply.
Images, videos and audio are available under their respective licenses.
