Wikipedia restricts AI writing; allows limited AI copyedits and translation

Wikipedia has published new rules restricting AI writing. The policy bans editors from using artificial intelligence (AI) to generate or rewrite article content, citing that text produced by large language models (LLMs) often conflicts with core Wikipedia content policies. The rules include narrow exemptions. Editors may use LLMs to suggest basic copyedits to their own writing, but only if the model does not introduce new, uncited information, and after human review. Wikipedia also allows LLM-assisted translation, but editors must follow the specified guidelines. Wikipedia did not detail penalties in the announcement. However, its LLM guidance warns that obviously disruptive AI misuse may be “struck or collapsed,” and repeated misuse could lead to a block or ban. The Wikimedia Foundation, which runs Wikipedia, reiterated that editorial policies remain human-centered and volunteer-driven. Separately, it has reported an ~8% year-over-year decline in human pageviews, attributing traffic pressure to search engines and AI chatbots. For traders, the broader angle in the article links AI accountability to blockchain. It argues that immutable, timestamped logs could help trace AI building, training, and usage decisions, potentially mitigating risks from deepfakes or unreliable LLM outputs. This part frames enterprise blockchain as an accountability layer for AI systems, though it is not a direct market-moving crypto protocol change. Overall, this is an information-governance update affecting AI writing workflows, not a crypto-native regulation event.