07.04.2026 12:27:00
Дата публикации
The popular online encyclopedia has officially approved new rules prohibiting editors from using large language models to write articles.
The decision was made after lengthy debates and is linked to risks of violating the platform’s key principles — verifiability and neutrality.
AI texts often contain “hallucinations”: fabricated facts, non‑existent references, and unreliable data.
Therefore, the use of ChatGPT and other models for generating encyclopedic content is now completely banned.
Exceptions are allowed only in two cases: translation of articles between languages and basic text editing.
Each change must be checked by a human and cannot add new information.
Previously, Wikipedia developed its own rules for detecting AI content, including signs such as clichés, excessive verbosity, and fake citations.
Wikipedia co‑founder Jimmy Wales called modern models “unreliable” and emphasized that they are not ready to replace human labor.
Experts warn: the ban may mark the beginning of a broader reassessment of AI’s role on platforms where information accuracy is valued.
This ban is also significant because AI is trained in part on numerous Wikipedia articles. Training will be of higher quality if it is based on content created by humans.
The decision was made after lengthy debates and is linked to risks of violating the platform’s key principles — verifiability and neutrality.
AI texts often contain “hallucinations”: fabricated facts, non‑existent references, and unreliable data.
Therefore, the use of ChatGPT and other models for generating encyclopedic content is now completely banned.
Exceptions are allowed only in two cases: translation of articles between languages and basic text editing.
Each change must be checked by a human and cannot add new information.
Previously, Wikipedia developed its own rules for detecting AI content, including signs such as clichés, excessive verbosity, and fake citations.
Wikipedia co‑founder Jimmy Wales called modern models “unreliable” and emphasized that they are not ready to replace human labor.
Experts warn: the ban may mark the beginning of a broader reassessment of AI’s role on platforms where information accuracy is valued.
This ban is also significant because AI is trained in part on numerous Wikipedia articles. Training will be of higher quality if it is based on content created by humans.
