Ars: Wikipedia is this perfect corpus of human-written text, full of human language about things that humans are interested in …
Harrison: And publicly licensed, too. It’s not perfect, but it’s a pretty clean data set in a lot of ways. And so I guess the existential threat is, if Wikipedia is not getting appropriately credited, will we forget about it and then stop maintaining it, and then it will no longer be as valuable a resource to both humans and AI?
Ars: You’re writing about volunteer editors, their personalities, their pride, motivations. Do you think it affects the editors behind Wikipedia if they get a sense they’re mostly writing for bots, for AI and Google to just summarize for searchers? Do you have any sense of whether editors care about the way they’re feeding people information?
Harrison: It’s a bit nuanced because Wikipedia editors aren’t the most egotistical people. They’re largely anonymous or writing behind pseudonyms. They’re not like TikTok stars who are trying to get their name out there. So in a way, they were never trying to claim credit for their edits and contributions to Wikipedia. So you could make the argument that because they’re relatively selfless, maybe it doesn’t matter if the data is going to AI or LLMs.
On the other hand, just from having talked to a lot of Wikipedia editors, they are very thoughtful of the reader. They envision a reader, and a lot of them are thinking about: Does this serve the reader? And I also think just a lot of them are put off by the idea that their work is just a big gift to Google, a for-profit company.
… It’s also a hard issue that I’m grappling with because what is the fix? Maybe it’s like, okay, these Big Tech companies are getting so much value from Wikipedia, maybe they should just donate more money to the Wikimedia Foundation. On the other hand, you don’t want the Big Tech companies to have any influence over the encyclopedia. So it’s not just that we want the big companies to reciprocate and influence the whole thing. It’s challenging.
Ars: It’s always fascinating to me how, by and large, every page is so much better than I think it will be. Really high-energy, maybe “truth optional” pages for topics like cryptocurrencies, it’s a small miracle that they’re as orderly and reasonable as they end up being. Do you have any thoughts on how, 24 years in now, Wikipedia has been able to maintain that kind of stability?
Harrison: Well, there are a lot of dedicated volunteers. If a page is getting a lot of eyeballs, like cryptocurrency, for example, that’s a page that people are going to be monitoring with some vigilance. With the mobile revolution, people are also watching edits on their phones, and can run back to their laptops to revert incorrect edits.
+ There are no comments
Add yours