[09:41:34] You may have read the robot-written article in The Guardian. In case you haven't, here it is: https://amp.theguardian.com/commentisfree/2020/sep/08/robot-wrote-this-article-gpt-3 [09:43:53] I mentioned it on Twitter and Lucas sent me this thing in German: https://uebermedien.de/47350/welt-und-t-online-bereiten-sich-auf-heisses-fussball-wochenende-vor/ [09:43:54] I know very little German, so I've read it with Google Translate (oh, the irony). Basically, it says that some German newspapers already publish some stories about sports based on available sports data and past results. If I understood correctly (the translation is fuzzy, and my understanding of football is even worse than my understanding of German), the articles are not quite relevant, because the games in question were canc [09:46:09] This article is quite acerbic and critical of this phenomenon of machine-written stories. My favorite part there was that it talked that it's for "das Klickvieh die Nutzer". "Klickvieh" ("click-cattle") is struck through and replaced with "Nutzer" (user). [09:47:04] Aaaanyway, in case you haven't guessed it already, what I'm getting at is that Abstract Wikipedia should be extra careful not to become pointless like that. [09:47:22] +1 (re @amire80: Aaaanyway, in case you haven't guessed it already, what I'm getting at is that Abstract Wikipedia should be extra careful not to become pointless like that.) [09:47:49] This depends on defining constraints for using Q-Objects [09:48:28] I'm quite sure that no one among its current developers and enthusiasts wants it to be pointless and to treat users as "cattle", but I hope a reminder won't hurt :) [09:49:18] Of course. The matter depends on many factors. (re @amire80: I'm quite sure that no one among its current developers and enthusiasts wants it to be pointless and to treat users as "cattle", but I hope a reminder won't hurt :)) [09:49:36] Several ones can be missed along this long way. [09:56:41] A curious thing there is that the article says that these robot-written stories were openly marked as such: "This article was created automatically by our partner Retresco based on game data". That's the kind of thing I referred to when I wrote that auto-generated text must be very clearly labeled, in a human-readable and machine-readable way (here: https://meta.wikimedia.org/wiki/Talk:Abstract_Wikipedia/Archive_1#Comparison_t [09:56:42] Now that I'm reading this story and the Guardian story, which was also clearly labeled, I'm realizing that while such marking is necessary, it's probably not enough. I'd really love this technology to go the extra mile with being ethical and making sure that everyone understands that a text was not written by humans, even people who don't notice that says "This article was created automatically". It's a challenge for real desi [09:56:54] A curious thing there is that the article says that these robot-written stories were openly marked as such: "This article was created automatically by our partner Retresco based on game data". That's the kind of thing I referred to when I wrote that auto-generated text must be very clearly labeled, in a human-readable and machine-readable way (here: https://meta.wikimedia.org/wiki/Talk:Abstract_Wikipedia/Archive_1#Comparison_t [09:56:55] Now that I'm reading this story and the Guardian story, which was also clearly labeled, I'm realizing that while such labeling is necessary, it's probably not enough. I'd really love this technology to go the extra mile with being ethical and making sure that everyone understands that a text was not written by humans, even people who don't notice that says "This article was created automatically". It's a challenge for real des [10:01:17] I wonder what Klickvieh would be in English... the only thing coming to mind is clickbait but that's not it [10:02:56] So, I don't know German, but it's struck out and replaced by "user" (Nutzer), so I guess it refers to the person who reads the article, and not to the clickbait, which is a phenomenon. [10:03:23] yeah, that's why it's not right :) [10:04:09] it seems to mean the group of people who are only wanted for their clicks on articles [10:04:33] and I can't think of a concise way of putting that in English, let alone a single word [10:04:43] The Google Translation into English is "click-cattle". The Google Translation into Russian has the same literal meaning, but the Russian word for "cattle" is also often used for people who are stupid, mindless, or evil. [10:06:16] I don't think that English has such a connotation, at least not as clearly as it is in Russian. I don't know which connotations does "Vieh" have in German, but maybe it's similar to Russian? In this context it's probably closer to the "mindless" sense than to the "evil" sense. [10:07:13] @lucaswerkmeister can you help us? :D [10:09:33] hm, clickbait targets? [10:09:36] yeah, I think German Vieh carries roughly the same connotation as the Russian word you mention [10:10:38] if a site treats its users as Klickvieh, it doesn’t care about them as users, it just wants its engagement counters to go up [10:11:06] yay I understood it [10:15:58] being just a number is a similar concept but that's also not a noun [10:18:07] This also vaguely reminds me of the issue of clearly marking old articles when they make the rounds again on social media. A prominent politician of our Green party sent https://twitter.com/RenateKuenast/status/1303306753895997440 yesterday: “CSU [center-right party, part of government since 2005, traditionally heads the ministry for transport] secretary general Söder [now head of CSU] demands ban on cars with inte [10:18:08] Now that I'm reading this story and the Guardian story, which was also clearly labeled, I'm realizing that while such labeling is necessary, it's probably not enough. I'd really love this technology to go the extra mile with being ethical and making sure that everyone understands that a text was not written by humans, even people who don't notice that says "This article was created automatically". It's a challenge for [10:20:26] I saw Guardian do this for some COVID-related articles that are just a weeks old. (re @lucaswerkmeister: This also vaguely reminds me of the issue of clearly marking old articles when they make the rounds again on social media. A prominent politician of our Green party sent this tweet yesterday: “CSU [center-right party, part of government since 2005, traditionally heads the ministry for transport] secretary general Söder [ [10:20:34] I saw Guardian do this for some COVID-related articles that are just a few weeks old. (re @lucaswerkmeister: This also vaguely reminds me of the issue of clearly marking old articles when they make the rounds again on social media. A prominent politician of our Green party sent this tweet yesterday: “CSU [center-right party, part of government since 2005, traditionally heads the ministry for transport] secretary general Söd [10:20:49] Like https://www.theguardian.com/world/2020/mar/14/anti-inflammatory-drugs-may-aggravate-coronavirus-infection [10:20:58] "This article is more than 5 months old" [11:09:32] "sheeple" is a somewhat related English term, people behaving like sheep, following the lead without much critical thinking. (re @Nikki: and I can't think of a concise way of putting that in English, let alone a single word) [11:14:40] Anyway, my point is that yes, there are people who are not so careful about what they read. Quite a lot of them probably. It's unfortunate, but it's a fact. Some publications abuse and exploit this by flooding them with clickbait, adware, disinformation, listicles, auto-created texts and videos, etc. [11:14:41] Wikipedia in general, and Abstract Wikipedia in particular could intentionally go the other way: recognize that some people are less careful, but design the products and the texts in a way that informs everyone responsibly, including such people. I'm not sure how to do it, but perhaps a real designer can think of good ideas. [11:14:44] But that is shifting the focus. "Klickvieh" is describing how the users are treated. "Sheeple" is descirbing how the people ara acting. (re @moebeus: "sheeple" is a somewhat related English term, people behaving like sheep, following the lead without much critical thinking.) [11:15:39] "somewhat related" [11:22:15] More like second cousin than sister, but ok 😉 (re @moebeus: "somewhat related") [11:25:32] true, a sister would be directly related 😉 [11:27:01] true, a sister would be directly closely related 😉 [11:28:22] +1 (re @amire80: Aaaanyway, in case you haven't guessed it already, what I'm getting at is that Abstract Wikipedia should be extra careful not to become pointless like that.)