Wikipedia Bans AI-Generated Content

Wikipedia has adopted a new rule that blocks the use of artificial intelligence to create or substantially rewrite material on the site. The change reflects concerns that large language models can conflict with the encyclopedia’s core standards and may introduce unreliable information.

The policy applies across Wikipedia and was described as taking effect recently. It was approved after a vote among the project’s volunteer editors, highlighting an ongoing internal debate about how far AI tools should be allowed to shape public-facing knowledge resources.

What changed, and what is still allowed

Under the updated approach, editors are prohibited from using AI systems to generate new text for articles or to rework existing passages in a way that amounts to rewriting. The stated rationale is that outputs from large language models can be misleading or fabricated, creating risks for accuracy and trust.

Two exceptions remain in place. AI tools can be used to translate content, and they can be used for limited copyediting that focuses on small wording improvements to an editor’s own writing.

Guardrails and the scale of the platform

Even where AI assistance is permitted, the policy sets conditions. Suggested edits must be checked by a person, and they must not add new information or change the intended meaning of the original text.

The decision lands on a platform with significant reach: the English-language Wikipedia has more than 7.1 million articles. That scale makes consistency and quality control central to how the encyclopedia maintains credibility across topics, languages, and regions.

Why it matters now

The move comes as AI tools become more common in how people search for and summarize information. The policy also responds to the wider shift in online discovery, where conversational AI products have become a competing gateway for general knowledge.

One data point cited in the broader discussion is that ChatGPT reportedly exceeded Wikipedia in monthly site visits last year. At the same time, the policy highlights a key limitation of current AI systems: they can produce confident-sounding statements that are incorrect, which can be difficult for readers to detect without verification.

Risks, uncertainties, and enforcement questions

While the rule draws a clear line against AI-authored or AI-rewritten article text, the practical challenge is enforcement at scale. The source material does not specify how Wikipedia will detect prohibited AI use, what penalties apply, or how disputes will be handled.

Another uncertainty is how the policy will evolve as AI tools improve and as editors seek productivity aids that do not compromise accuracy. For now, Wikipedia’s approach prioritizes human review and limits AI to translation and small, non-substantive edits.