We often write, somewhat stressed, that AI is a threat to us and that the accusation us having a bias in our articles hurts our credibility.
I think we should look at our reality with another view. We have by now a huge number of very good articles on entities being typical encyclopedian, like oxygen or Stockholm. AI will not create better articles and no one calls them biased. And we have a huge number of very dull articles, like all the lakes in Canada, where AI can not produce anything as good, and no one calls them biased.
For the one called biased I have found they represent about 0,01% of all, and when I talk with readers, they anyway say they do not trust Wikipedia (based on the subject, not the actual content) on these entries and the proportion of readers finding fact on these things in Wikipedia compared with other sources are l relative low (specially compared with the typical encyclopedian entries). Why do we spend so much energy on these few articles that anyway gives us very little or no/negative credit? I half seriously on my home wiki is trying to launch the concept of the "dull Wikipedia". Where we write very short articles, with only dull basics, of these controversial subjects and also avoid the juicy content of of individuals even if it takes a lot of volume in media.
Why not concentrate and be proud of the thing we are best in. Noncontroversial standard encyclodedan entries, including marginal subjects? And leave the controversial subjects to the media, and the subjects where AI excels in to that realm? And forget concept like click rate. And be proud to the good we are doing
Anders
Because for many of us it is funny writing about them.
Writing about dull stuff just to improve content sounds like work, and work I/we already have in another part of the day, so we usually write about what mostly pleases us, and that often means pop culture, niche of the niche of the niche subjects and controversial stuff. :)
Best, Paulo
Anders Wennersten mail@anderswennersten.se escreveu (terça, 15/10/2024 à(s) 11:08):
We often write, somewhat stressed, that AI is a threat to us and that the accusation us having a bias in our articles hurts our credibility.
I think we should look at our reality with another view. We have by now a huge number of very good articles on entities being typical encyclopedian, like oxygen or Stockholm. AI will not create better articles and no one calls them biased. And we have a huge number of very dull articles, like all the lakes in Canada, where AI can not produce anything as good, and no one calls them biased.
For the one called biased I have found they represent about 0,01% of all, and when I talk with readers, they anyway say they do not trust Wikipedia (based on the subject, not the actual content) on these entries and the proportion of readers finding fact on these things in Wikipedia compared with other sources are l relative low (specially compared with the typical encyclopedian entries). Why do we spend so much energy on these few articles that anyway gives us very little or no/negative credit? I half seriously on my home wiki is trying to launch the concept of the "dull Wikipedia". Where we write very short articles, with only dull basics, of these controversial subjects and also avoid the juicy content of of individuals even if it takes a lot of volume in media.
Why not concentrate and be proud of the thing we are best in. Noncontroversial standard encyclodedan entries, including marginal subjects? And leave the controversial subjects to the media, and the subjects where AI excels in to that realm? And forget concept like click rate. And be proud to the good we are doing
Anders
Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l Public archives at https://lists.wikimedia.org/hyperkitty/list/wikimedia-l@lists.wikimedia.org/... To unsubscribe send an email to wikimedia-l-leave@lists.wikimedia.org
Kaya
AI doesnt excel in anything, it can only be a copy of the sources it uses, that does include Wikipedia regardless of which language project it is.
Our benefit is being "neutral" in both how we write and what we write about, like Paulo just said its fun to dig into various topics including controversial stuff. Science can be as if more controversial than people, politics, or religion but our goal is to "Share the sum of all knowledge" to do less is a disservice to those seeking genuine unbias content.
I'd much prefer a world where Wikipedia is the only source, but if AI is going to exist and what we do makes it regurgitate better information then that's good.
On Tue, 15 Oct 2024 at 18:08, Anders Wennersten mail@anderswennersten.se wrote:
We often write, somewhat stressed, that AI is a threat to us and that the accusation us having a bias in our articles hurts our credibility.
I think we should look at our reality with another view. We have by now a huge number of very good articles on entities being typical encyclopedian, like oxygen or Stockholm. AI will not create better articles and no one calls them biased. And we have a huge number of very dull articles, like all the lakes in Canada, where AI can not produce anything as good, and no one calls them biased.
For the one called biased I have found they represent about 0,01% of all, and when I talk with readers, they anyway say they do not trust Wikipedia (based on the subject, not the actual content) on these entries and the proportion of readers finding fact on these things in Wikipedia compared with other sources are l relative low (specially compared with the typical encyclopedian entries). Why do we spend so much energy on these few articles that anyway gives us very little or no/negative credit? I half seriously on my home wiki is trying to launch the concept of the "dull Wikipedia". Where we write very short articles, with only dull basics, of these controversial subjects and also avoid the juicy content of of individuals even if it takes a lot of volume in media.
Why not concentrate and be proud of the thing we are best in. Noncontroversial standard encyclodedan entries, including marginal subjects? And leave the controversial subjects to the media, and the subjects where AI excels in to that realm? And forget concept like click rate. And be proud to the good we are doing
Anders
Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l Public archives at https://lists.wikimedia.org/hyperkitty/list/wikimedia-l@lists.wikimedia.org/... To unsubscribe send an email to wikimedia-l-leave@lists.wikimedia.org
A nice thing is that the AI I've been using (mostly ChatGPT) seems to understand well how Wikipedia works, and often displays a kind of veneration for our work (it knows where it gets their stuff), often open and shamelessly displays Wikipedia as "source" and in general shows us in a very positive light to whoever asks. I believe it may have a positive effect in outreach and engagement.
Paulo
Gnangarra gnangarra@gmail.com escreveu (terça, 15/10/2024 à(s) 11:44):
Kaya
AI doesnt excel in anything, it can only be a copy of the sources it uses, that does include Wikipedia regardless of which language project it is.
Our benefit is being "neutral" in both how we write and what we write about, like Paulo just said its fun to dig into various topics including controversial stuff. Science can be as if more controversial than people, politics, or religion but our goal is to "Share the sum of all knowledge" to do less is a disservice to those seeking genuine unbias content.
I'd much prefer a world where Wikipedia is the only source, but if AI is going to exist and what we do makes it regurgitate better information then that's good.
On Tue, 15 Oct 2024 at 18:08, Anders Wennersten mail@anderswennersten.se wrote:
We often write, somewhat stressed, that AI is a threat to us and that the accusation us having a bias in our articles hurts our credibility.
I think we should look at our reality with another view. We have by now a huge number of very good articles on entities being typical encyclopedian, like oxygen or Stockholm. AI will not create better articles and no one calls them biased. And we have a huge number of very dull articles, like all the lakes in Canada, where AI can not produce anything as good, and no one calls them biased.
For the one called biased I have found they represent about 0,01% of all, and when I talk with readers, they anyway say they do not trust Wikipedia (based on the subject, not the actual content) on these entries and the proportion of readers finding fact on these things in Wikipedia compared with other sources are l relative low (specially compared with the typical encyclopedian entries). Why do we spend so much energy on these few articles that anyway gives us very little or no/negative credit? I half seriously on my home wiki is trying to launch the concept of the "dull Wikipedia". Where we write very short articles, with only dull basics, of these controversial subjects and also avoid the juicy content of of individuals even if it takes a lot of volume in media.
Why not concentrate and be proud of the thing we are best in. Noncontroversial standard encyclodedan entries, including marginal subjects? And leave the controversial subjects to the media, and the subjects where AI excels in to that realm? And forget concept like click rate. And be proud to the good we are doing
Anders
Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l Public archives at https://lists.wikimedia.org/hyperkitty/list/wikimedia-l@lists.wikimedia.org/... To unsubscribe send an email to wikimedia-l-leave@lists.wikimedia.org
-- Boodarwun Gnangarra 'ngany dabakarn koorliny arn boodjera dardon nlangan Nyungar koortabodjar'
Wikimedia-l mailing list -- wikimedia-l@lists.wikimedia.org, guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l Public archives at https://lists.wikimedia.org/hyperkitty/list/wikimedia-l@lists.wikimedia.org/... To unsubscribe send an email to wikimedia-l-leave@lists.wikimedia.org
wikimedia-l@lists.wikimedia.org