Wikipedia bans AI: a bulwark against digital misinformation

⚡

Key Takeaways

1Wikipedia prohibits the use of generative AIs in its articles, citing violations of its editorial policies.

2The decision follows months of battling biased and erroneous content generated by LLMs like ChatGPT.

3Some exceptions remain for minor tasks, but human oversight is still required to ensure accuracy.

💡Why it matters — This ban aims to preserve the reliability of Wikipedia, which is essential in a digital landscape overwhelmed by AI.

Wikipedia Takes a Stand Against Generative AI

In a bold move, Wikipedia has decided to ban the use of advanced language models, or LLMs, in the writing of its articles. This decision reflects the online encyclopedia's firm commitment to protecting the integrity of its content in the face of the rise of artificial intelligence. Since its inception 25 years ago, Wikipedia has enabled millions of contributors to share knowledge, provided they rely on trustworthy sources. Today, the encyclopedia is confronted with a significant challenge: the invasion of generative AI that threatens the quality and accuracy of its articles.

A Community Decision

The Wikipedia community recently voted to prohibit the use of generative AI tools such as ChatGPT and Gemini in the creation or modification of articles. This decision is based on the observation that these technologies often violate the fundamental principles of the encyclopedia, particularly regarding neutrality and verifiability. Wikipedia contributors have already spent months battling AI-generated articles, even forming a dedicated group for their removal. Last year, Wikipedia had already eliminated an automatic summary feature, citing the risks of factual errors and biases in interpretations.

Exceptions Under Scrutiny

Despite this ban, Wikipedia still allows the use of AI for specific tasks, such as minor corrections, provided that an administrator validates the content. Similarly, the translation of articles from one language version to another can utilize AI, but only if the editor is sufficiently proficient in both languages to ensure the accuracy of the translation. These exceptions demonstrate that Wikipedia acknowledges the potential utility of AI while insisting on the necessity of rigorous human oversight.

A Detection Challenge

Detecting the use of AI in articles remains a major challenge for Wikipedia. The encyclopedia admits that some human writers can naturally produce texts that resemble those generated by LLMs, making it difficult to identify problematic content based solely on style. Moderators must therefore rely on assessing compliance with editorial policies and reviewing the contribution history of editors. However, this approach is not foolproof, especially for pages that do not receive constant monitoring.

Variable Rules by Language

It is important to note that this ban primarily applies to the English version of Wikipedia, which is the most consulted. Each language version of the encyclopedia operates autonomously and can establish its own rules. For example, the Spanish version of Wikipedia has adopted an even stricter stance by completely banning the use of LLMs, with no exceptions. This diversity of rules reflects the different approaches of linguistic communities in response to the impact of AI on content creation.