Wikipedia:AI generated content
![]() | This is a suggested policy, guideline, or method for Simple English Wikipedia. This page is not a policy, guideline, or method yet, but may be soon. Look at its talk page for more information. |
![]() | This page in a nutshell: AI generated content must adhere to complexity guidelines, and not violate copyright laws. Editors are solely responsible for the content they post |
With the new artificial intelligence models, it is now possible to generate text for a Wikipedia page with just a few keystrokes asking for a page to be written on any subject. This can make it very easy to create new Simple English Wikipedia content, but it has drawbacks. All such content must follow the rules below in order to comply with the policies and guidelines on Simple Wikipedia as well as copyright law. Pages that include AI content must include the template {{AI-notice}}.
Complexity and completeness
[change source]As with all content, AI-generated content must:
- Adhere to the Manual of Style
- Contain reliable sources to have proper citations
- Use simple language so as to comply with complexity guidelines such as a simple introduction
Any AI generated content that is too complex could be deleted at RfD. Not only does the content need to be simple, it must be accurate and complete enough to explain the subject to the reader or it can also be deleted.
Asking an AI to "write a Wikipedia article" can sometimes cause the output to be outright fabrication, complete with false references. It may be biased, may libel living people, or may violate copyrights. Thus, all text generated by AIs should be verified by editors before use in articles.
Copyright
[change source]In the United States, the law says that a person generating AI content is the copyright holder of that content. Therefore, the editor creating AI content must set such content as public domain or under the latest Creative Commons license. This will be automatically agreed to upon publishing such content on Simple English Wikipedia.
AI-generated content may include copyrighted information that should not be included without a proper citation to the original text. Again, this is the responsibility of the editor.