2 months ago
English Wikipedia has banned nan usage of generative AI erstwhile penning aliases rewriting articles. The level says it came to this determination because utilizing AI to whip up transcript "often violates respective of Wikipedia's halfway contented policies."
There are a mates of insignificant exceptions. Editors tin usage ample connection models (LLMs) to refine their ain writing, but only if nan transcript is checked for accuracy. The argumentation states that this is because LLMs "can spell beyond what you inquire of them and alteration nan meaning of nan matter specified that it is not supported by nan sources cited."
Editors tin besides usage LLMs to assistance pinch connection translation. However, they must beryllium fluent capable successful some languages to drawback errors. Once again, nan accusation must beryllium checked for inaccuracies.
"My genuine dream is that this tin spark a broader change. Empower communities connected different platforms, and spot this go a grassroots activity of users deciding whether AI should beryllium invited successful their communities, and to what extent," Wikipedia administrator Chaotic Enby wrote. The administrator besides called nan argumentation a "pushback against enshittification and nan forceful push of AI by truthful galore companies successful these past fewer years."
There is 1 point worthy noting. Wikipedia is not a monolith. Each Wikipedia tract has its ain independent rules and editing teams. Some whitethorn determine to clasp LLMs. However, others whitethorn spell moreover further. Spanish Wikipedia, for instance, has afloat banned nan usage of LLMs, with nary exceptions for refinement aliases translation.
Also, identifying matter written by LLMs is not an nonstop subject truthful Wikipedia's quality moderators could miss immoderate spots of slop each now and again. This is much apt connected pages pinch little predominant moderation.
