Wikipedia:Artificial intelligence#Article text generation

{{Further|meta:Artificial intelligence}}

AI-related efforts on Wikipedia include but are not limited to:

The Objective Revision Evaluation Service (ORES) was started in 2015 as a project of the Wikimedia Foundation, and provides a revision score against machine learning models that have been trained in order to report article quality or vandalism. This is used in tools such as ClueBot NG to help immediately revert vandalism, or in evaluation tools like the Program and Events Dashboard to measure the outcomes of classwork, edit-a-thons, or organized editing campaigns.

Guidance can be found at Help:Translation#English Wikipedia policy requirements. There is a Content Translation Tool used across Wikimedia projects that can use the output of machine translation from one Wikipedia article to another, using services like Google Translate. However, on the English Wikipedia, it currently states that "machine translation is disabled for all users and this tool is limited to extended confirmed editors." As a result, only manual translation on the English Wikipedia is supported by the tool, though some users have used translation to Simple English as a workaround. Relatedly, there is a section of the Help:Translation page with the broad advice: "avoid machine translations." However, this guidance was last edited in 2016, and the state of the art for machine translation has advanced significantly since then, which may merit a re-examination of that advice.

{{See also|Wikipedia:Large language models}}

The explosion of interest in ChatGPT in 2022 has led to increased curiosity in using generative AI to help compose Wikipedia articles. The status of machine-generated text from tools such as ChatGPT is generally accepted to be public domain, so the copyright issues are not a blocker to using the generated text from a legal standpoint. These issues are generally governed by Help:Adding open license text to Wikipedia#Converting and adding open license text to Wikipedia, which advises to make sure content is adjusted for style and that reliable sources are used. Conversations on the Village Pump and in some test articles (i.e. Artwork title) have noted positive aspects of machine generated text, but a serious warning that content must be checked for facts and accuracy and never used straight from ChatGPT.

A good general page looking at the issues can be found at: Wikipedia:Using neural network language models on Wikipedia.

In 2022, a major community discussion took place on Village Pump (policy) found at: Wikipedia:Village pump (policy)/Archive 179#Wikipedia response to chatbot-generated content

Some user experiences can be found here:

• Talk:Artwork title
• User:JPxG/LLM demonstration
• User:Fuzheado/ChatGPT - also: experiments with generating Wikidata Quickstatements from fuzzy date descriptions
• User:DraconicDark/ChatGPT
• User:BrokenSegue - Wikidata:Wwwyzzerdd and [https://wikidata-game.toolforge.org/distributed/#game=84 Psychiq] Wikidata game that uses distilBERT and ML, analyzing Wikipedia categories.

Image metadata – There have been efforts from GLAM institutions to help supplement image keyword data with machine learning efforts. Among them include:

• Computer aided tagging Started in 2019, "The computer-aided tagging tool is a feature in development by the Structured Data on Commons team to assist community members in identifying and labeling depicts statements for Commons files." See: :c:Commons:Structured data/Computer-aided tagging
• Metropolitan Museum of Art Tagging - This project used Met Museum tagging info to train a machine learning system to help predict new "depiction" recommendations for Wikidata. This resulted in a new Wikidata Game that helped add more than 4,000 new depiction (P180) statements to Wikidata. See the Met Museum blog post by Andrew Lih: "Combining AI and Human Judgment to Build Knowledge about Art on a Global Scale," March 4, 2019, [https://www.metmuseum.org/blogs/now-at-the-met/2019/wikipedia-art-and-ai]

Image generation

• Wikimedia Commons and AI generated media
• AI images and German Wikipedia, results of a meeting
• [https://www.youtube.com/watch?v=tvyrz4Y44iU A Battle for Reality], video essay on AI images and Wikipedia
• Wikimedia Commons AI, a rejected proposal for a new Wikimedia sister project aimed at establishing a clear distinction between human-generated content and content produced by artificial intelligence
• The four categories, an idea about dividing all images uploaded to Wikimedia Commons in one of four categories

{{Shortcut|WP:AIPOLICY}}

At the onset of the 2020s AI boom, Wikipedia's existing content policies already addressed many emerging AI-related concerns that would prompt other platforms and organizations to adopt a dedicated new policy; consequently, Wikipedia has no single "AI-generated content policy", "AI content guideline", et cetera. {{slink|Wikipedia:Large language models|Risks and relevant policies}} (essay) aims to explain how the broad core content policies and the copyrights policy interact with the use of AI tools, mostly in the domain of text. However, there exist disparate portions of policies and guidelines which are specifically and explicitly about AI-generated content, and they are listed here as follows ({{As of|2025|5|bare=y}}):

• {{slink|Wikipedia:Image use policy|AI-generated images}} (WP:AIIMAGES), a policy section against the use of images wholly generated by AI
• referred to in {{slink|Wikipedia:No original research|Original images}} (WP:IMAGEOR)
• portion of {{slink|Wikipedia:Biographies of living persons|Images}} (WP:AIIMGBLP), a policy norm against the use of AI-generated images to depict subjects of BLPs
• portion of {{slink|Wikipedia:Public domain|Works ineligible for copyright protection}} (WP:NONCREATIVE), a guideline paragraph recounting the legal principle that works created by machines are not copyrightable (generally valid as of 2025{{Cite web |title=Copyright and Artificial Intelligence |url=https://copyright.gov/ai/ |website=United States Copyright Office |access-date=April 9, 2025 }})
• portion of {{slink|Wikipedia:Manual of Style/Images|Editing images}} (MOS:AIUPSCALE), a MoS norm against the use of AI upscaling software
• {{slink|Wikipedia:Reliable sources|Sources produced by machine learning}} (WP:RSML), a guideline section against citing AI-generated content as sources
• referred to in {{slink|Wikipedia:Reliable sources/Perennial sources#Large language models}} (WP:RSPCHATGPT), an information page
• {{slink|Wikipedia:Talk page guidelines#LLM-generated}} (WP:AITALK), a guideline entry allowing the striking or collapsing of comments that are obviously generated by an LLM or similar AI technology

The following are not policies or guidelines, but still have some significance in this context:

• an entry in {{slink|Wikipedia:Drafts|Reasons to move an article to draftspace}} (WP:DRAFTREASON), an explanatory essay about policies on editing and deletion
• {{slink|Help:Translation|Automatic translation}} (WP:MACHINE)—concerns neural machine translation (i.e., "AI translation")
• portion of {{slink|Wikipedia:Guide to appealing blocks|Composing your request to be unblocked}} (WP:NICETRY), an explanatory essay stating that unblock requests that appear to be written using AI are likely to be summarily rejected

• Wikipedia:Large language models, an essay on using LLMs (textual generative AI) to produce or modify content on Wikipedia
• Wikipedia:Computer-generated content, a draft of a proposed policy on using computer-generated content in general on Wikipedia
• Wikipedia:WikiProject AI Cleanup, a group of editors focusing on the issue of non-policy-compliant LLM-originated content
• Wikipedia:Using neural network language models on Wikipedia, an essay about large language models specifically
• Artwork title, a surviving article initially developed from raw LLM output (before this page had been developed)
• m:Research:Implications of ChatGPT for knowledge integrity on Wikipedia, an ongoing (as of July 2023) Wikimedia research project
• m:Wikilegal/Copyright Analysis of ChatGPT
• Initial version of Artwork title, a surviving article developed from raw LLM output
• [https://slate.com/technology/2023/01/chatgpt-wikipedia-articles.html Should ChatGPT Be Used to Write Wikipedia Articles?], a Slate article which largely deals with the history and implications of 'Artwork title'
• Artificial intelligence in Wikimedia projects

• {{cite web|url=https://www.metmuseum.org/blogs/now-at-the-met/2019/wikipedia-art-and-ai|title=

Combining AI and Human Judgment to Build Knowledge about Art on a Global Scale|first=Andrew|last=Lih|date=March 4, 2019|publisher=Metropolitan Museum of Art}}

• {{cite web|url=https://wikimediafoundation.org/news/2019/07/18/designing-ethically-with-ai-how-wikimedia-can-harness-machine-learning-in-a-responsible-and-human-centered-way/|title=Designing ethically with AI: How Wikimedia can harness machine learning in a responsible and human-centered way|last=Morgan|first=Jonathan T.|date=18 July 2019|publisher=WIkimedia Foundation}}
• {{cite web|url=https://wikimediafoundation.org/news/2018/03/14/machine-learning-visually-enriching-wikidata/|title=How we're using machine learning to visually enrich Wikidata|last=Redi|first=Miriam|date=14 March 2018|publisher=Wikimedia Foundation}}
• :meta:Research:Ethical and human-centered AI

• User:JPxG/LLM demonstration (wikitext markup, table rotation, reference analysis, article improvement suggestions, plot summarization, reference- and infobox-based expansion, proseline repair, uncited text tagging, table formatting and color schemes)
• User:JPxG/LLM demonstration 2 (suggestions for article improvement, explanations of unclear maintenance templates based on article text)
• User:Fuzheado/ChatGPT (PyWikiBot code, writing from scratch, Wikidata parsing, CSV parsing)
• User:DraconicDark/ChatGPT (lead expansion)
• Wikipedia:Using neural network language models on Wikipedia/Transcripts (showcases several actual mainspace LLM-assisted copyedits)
• User:WeatherWriter/LLM Experiment 1 (identifying sourced and unsourced information)
• User:WeatherWriter/LLM Experiment 2 (identifying sourced and unsourced information, including a non-English source)
• User:WeatherWriter/LLM Experiment 3 (identifying sourced and unsourced information, only six of seven tests successful)
• Wikipedia:Articles for deletion/ChatGPT and Wikipedia:Articles for deletion/Planet of the Apes (humorous April Fools' nominations generated almost entirely by large language models).

{{reflist}}