In summary:
- Wikipedia has banned the use of large language models for generating or rewriting article content on its English-language platform, according to PCWorld.
- The new policy allows AI for basic text editing and translations, but requires mandatory human review and verification of accuracy.
- This cautious approach prioritizes human oversight while other Wikipedia language versions may establish their own separate AI content rules.
After much debate, Wikipedia has now taken a stance on AI-generated content on its platform: “the use of LLMs to generate or rewrite article content is prohibited.” So says the internal policy page, although the declaration does come with a few exceptions.
Wikipedia editors are permitted to use AI services for basic editing of text they’ve written themselves. Any AI-altered text must be reviewed by humans, both to ensure that the AI model hasn’t added its own material and that the core meaning of the text wasn’t changed.
Wikipedia is also allowing AI for translations. AI services may be used to produce an initial version of a translated Wikipedia article, but the translating editor must themselves be sufficiently proficient in both (original and translated) languages to be able to check that the translation is accurate and without errors.
These new rules apply only to the English-language Wikipedia. Wikipedia editors for other languages may come up with their own rules and guidelines for using AI in their articles.
This article originally appeared on our sister publication PC för Alla and was translated and localized from Swedish.