Wikipedia bans the use of AI models for the creation of articles, as the text that is produced by large language models (LLMs) often violates basic content policies of Wikipedia.
One policy that is violated is that of non-original content, that is articles that are based strictly on sources, something that AI often does not do.
This move reflects the shift of the climate among the editors of Wikipedia towards AI, from “cautious optimism… to real concern”, as one of them stated to 404 Media, with reports for problems that are related to AI increasing.
Wikipedia is a free online encyclopedia, it was founded in 2001 and has millions of published articles in many languages worldwide.
Somewhat ironically, Wikipedia itself constitutes a useful model for the adoption of AI, as noted recently by a researcher of Google DeepMind: a technologically driven source of knowledge, which was initially treated with skepticism as unreliable, but can constitute a powerful tool provided that it is used for the finding of primary sources and not for direct reference.
The new policy, which was approved with an overwhelming majority of 40 to 2 among the editors, will allow editors to use LLMs to suggest basic corrections to their own text, which can be incorporated after human review.
“Attention is required, as LLMs may exceed what is asked of them and change the meaning of the text in a way that is not supported by the sources that are cited,” the policy points out.
Source: 404media
What is fyi.news?
