36
votes
How Wikipedia’s editors keep the online encyclopedia human-written
Link information
This data is scraped automatically and may be incorrect.
- Title
- Inside Wikipedia's Cleanup Effort Against AI Text
- Authors
- Eliza Strickland
- Published
- Oct 21 2026
- Word count
- 1163 words
Wikipedia is a very near and dear project to me, I've donated to them for many years now even though they might not need it cause when you donate to them you get routed to a cool page with tshirts and swag you can buy and I enjoy that a lot, plus, donating is good for the soul.
Being a place that I can rely on to not be AI generated is super important to me, and I'm glad it's important to them, because it's not important anywhere else in my life, except maybe, here.
I sometimes review article drafts submitted for publishing on Wikipedia, and there are so many new accounts that just submit blatantly AI generated articles on random topics. I have no idea why they do that. When it's an article about a company I at least understand that they're trying to advertise themselves on Wikipedia. But why would someone submit a shitty AI-generated article on an obscure bug species? What do they gain from it???
Another unfortunately pretty frequent occurrence is that someone under the influence of AI psychosis believes that they just came up with some revolutionary new scientific idea, and then submit a 10000 word long AI article which must seem brilliant to them but is actually a completely incomprehensible schizo word salad.
I saw an article or video (don't remember which, now) about a guy who's published dozens or hundreds of books on a wide range of subjects on Amazon, all "co-written" with an LLM. He claims to feel a true sense of accomplishment for the volume of books he's produced, and I believe him having seen his enormous shelf of his own books (he always buys a physical copy of his own books as soon as they're on Amazon).
I think some people genuinely believe they're contributing something of value, that they have the idea for what they want to see but not the expertise to "wordsmith" it, so they get "a little help," in their minds equivalent to spell-check only a bit moreso. Unfortunately they also tend to lack the expertise to know whether the output is any good, and they are unaware that (in most cases) being able to speak clearly on a topic is one of the foundational demonstrations of one's proficiency.
I know there are exceptions, savants who are the best in a topic but just can't get the words out. But they are a vanishingly small minority, and the majority of people doing this have a bad case of "but my opinion matters" regardless of its uninformed nature.
Ironically, speaking fluently on almost any topic, to the degree most non-experts in the subject can't tell it's BS, is what LLMs are best at.
The chatbot tells the author that their ideas are brilliant and "proves" it by spitting out a fluent expansion on whatever they prompted.
People who previously didn't feel like they had any talent for creative or intellectual output suddenly feel like they're amazing at both. They just needed a little help with the words!
It baffles me on a human level too. It's compounded by the fact the AI bot is probably trained on what Wikipedia already has. The chance the AI adds something genuinely new and well sourced is probably pretty small.
The only real reason I can think of is how bot accounts frequently operate: build legitimacy, then astroturf.
I can think of plenty of reasons when it comes to political topics, up to and including economic policies, history, and all that sort of stuff, but not so much when it comes to simple factual articles like your bug example.
Yeah, but there isn't any sort of "karma" or account reputation on Wikipedia, so I don't know what they're trying to achieve. The only somewhat relevant concept is an "autoconfirmed" account which lets you directly create articles without first going through the review queue, but for that you just need 10 edits of any kind (which IMO is way way too low of a bar, but unfortunately many other editors believe that the review queue is too slow and bureaucratic and discourages people from submitting articles, so they don't want to raise it)
Maybe they’re doing what accounts on reddit do, posting believable posts to legitimize their account so that later on they can make a sneaky update to their company’s wiki
Reading this is making me want to update my company's Wikipedia entry with a list of outages attributable to AI usage...
Just yesterday, me and my coworkers got into a conversation about the world's oldest university. One guy brings up the University of Timbuktu, which I look up on WP and try to correct that it's not a single, organised university - it's a collective name for all the research and teaching that went on in the community. Before I do they jump down my throat about using Wikipedia because of how deeply unreliable it is and how all their teachers told them to never use it as a source. "But", I say, "all your research is from asking ChatGPT!".
"Yeah," they say, "but all of its output comes with references!"
"Ah,", I reply, "and the huge list of references at the end of ever single WP article, they don't count. Right?"
We are so doomed....
Interesting, whenever people tell me Wikipedia isn’t a reputable source they’re usually trying to prove some obviously wrong thing like antivaxxing and when pressed they’ll tell me the whole internet is a lie but somehow they have the facts cause theyre special or something
I do think there will be value in maintaining a strictly human made encyclopedia. It's like how knitting can be done by machine and mass produced but there is still value and art to hand knitting your own projects.
It's significance may change over time, but I appreciate that Wikipedia is trying to keep clean of AI generated text and maintain truth.