Editors using the visual editor will now see a warning when they paste text that is likely to have initially been copied from one of five external generative AI chatbots into Wikipedia. This feature is known as LLM Paste Check. It is a type of Edit Check that is community configurable. The check looks for markers in the copied HTML and is not using content watermark detection. Edits will also be tagged so that patrollers can review them more closely. The feature helps newcomers add content that respects Wikipedia’s policies and makes it easier for experienced editors to identify problematic edits. Communities with local policies about LLM usage can adjust the wording and policy that is linked. The team apologizes that this was rolled out without a delay for translations.
It will 'write articles' just fine. Trouble is, it will write inaccurate articles and simply make up facts and citations whenever it's missing any.
I'd hate to be the wikipedia editor who has to come in after that and clean all that up, fact-checking each claim, checking to make sure all the references actually exist and actually say what the text claims they say. It would be super-tedious.
My point was more of, what other site could find a more useful purpose for AI writing on earth? We're being told, and have been told that AI can do all these things, right? But not even Wikipedia can realistically find a good use for it in the one thing AI is supposed to have 'mastered'. It just made me reflect.
what other site could find a more useful purpose for AI writing on earth?
It's great for generating fake product reviews!
Oh, and fake engagement for social media. Fantastic for that.
LLMs have zero accountability when it comes to fact checking. They have been shown time and time again to just make shit up.
Factual sounding bullshit is absolutely not good, and definitely not "better than humans" for a system that aims to be a fact repository. Sure, they can be a tool to help with writing, but only if the author 100% knows what they're talking about and just needs help with editing.
People who paste raw LLM output into Wikipedia generally do not know what they're talking about as well as they think they do.
Also, it is not "downright banned" - it gets a warning and flagged for review because the frequency at which LLM copy-paste is bullshit is too high.
@lemmy.world
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
go to feed...
@lemmy.world
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
go to feed...
The goalposts on AI have moved so far I almost forgot that writing even simple articles was something it was supposed to do pretty well or better than humans. Now it's downright banned.
save