WEBVTT

1
00:00:00.001 --> 00:00:00.470
So

2
00:00:00.890 --> 00:00:06.440
by 30 Friday evening, Friday afternoon, no Friday morning

3
00:00:06.980 --> 00:00:17.559
and Friday early morning, I think, on the other side of the world. Good to see you all here, very happy to have everybody around and

4
00:00:17.610 --> 00:00:42.929
we we we were excited, both of us. I want to thank Adola first of all for organizing and hosting this session. I'm just here, sharing and talking everything together. he did all the work. I didn't. So thank you. And welcome to this great session about algorithms and society topic close to my heart, so I'm excited to hear the presentations today. but before that, let me start with a acknowledgment of country

5
00:00:43.090 --> 00:00:58.740
on behalf of everyone everyone gathered. I pay respect to the spirit of the country. Here in Nam, where we are meeting, I acknowledged the past and present culmination. Bonjeri voiced as the sovereign people. Here the cultural knowledge holders.

6
00:00:59.630 --> 00:01:27.149
we maintain and promote the norm aboriginal law, customs and custodianship, responsibilities for country and the people that country continues to sustain generation after generation into the future. For all of us I recognize the important contribution that first nation thinkers have made to sociology, including challenges to our dominant way of thinking. We thank all first nation delegates in the room today for their contributions to knowledge and the discipline.

7
00:01:28.150 --> 00:01:40.829
and with that I would like to, I think. Yes, I would like to invite the first presenter, Karina to come up. let me start your presentation.

8
00:01:42.950 --> 00:02:10.290
It's also Friday evening for this system. Great? Thank you for introducing me. yeah, I'm Krina, and I'm a Phd. Candidate in political sociology at the University of Adelaide in South Australia, and my presentation is called hamming the algorithm. And I wish I would have known how to do that today because I was trying to get tailors with tickets in Melbourne, but that didn't work out.

9
00:02:11.220 --> 00:02:29.700
they are so. I were presenting some preliminary findings from interviews I conducted with social media influence as to share political content and focus on how they position themselves and their content within the algorithms of social media platforms, employing a body Asian perspective.

10
00:02:30.110 --> 00:02:44.280
 just briefly, this research is part of my Ph. D. Project, where I'm generally looking at the shift towards political content shared that influences. So it users users who are more active, more popular, and thought to be more influential than others.

11
00:02:44.480 --> 00:02:50.220
and how this is shaping political views and the socialization of followers and other users.

12
00:02:50.560 --> 00:03:02.509
But today's presentation is focusing on this like outer square that you see here and is guided by the question of how influences interpret algorithmic decisions and express agency when they navigate the space.

13
00:03:03.710 --> 00:03:17.759
And to make this a bit more visual, here are some screenshots of my Instagram account feed. And I guess we all know what that is. But the Instagram feed is the first page that is displayed when you open up the app, which is kind of comparable to the Facebook timeline, for example.

14
00:03:17.760 --> 00:03:38.089
And what we see here is my algorithmically curated feed, including on the left some German news pages that I follow course, adds the supermarket here in Australia, if you've been to to it yet some Phd humor travel, content suggestions for you pages that the algorithm presents to me is presumably interesting. On the top right, for example.

15
00:03:38.670 --> 00:03:39.880
And

16
00:03:40.980 --> 00:04:03.639
here are 3 examples of posts that my participants actually shared regarding a change in abortion or in Germany. And the main question is, How do How do social media and Francis to share? Such content negotiate the visibility of this content emits platforms, affordances and algorithms, and so to to be displayed on the feed or to be suggested by the algorithm.

17
00:04:05.380 --> 00:04:22.710
and in the next, like, less than 15 min now I will talk about the theoretical background of algorithms on social media platforms and present another way of looking at them, employing a booze in perspective that moves away from a technological, deterministic view and takes into account human agency and algorithmic functioning.

18
00:04:23.380 --> 00:04:30.220
And I will provide a concise overview of the methods applied, and present preliminary findings from the interviews I conducted with influences

19
00:04:32.370 --> 00:04:43.220
so social media platforms have been seen as information intermediaries, or gatekeepers that afford the architecture and underlying algorithms that determine how information on them is presented.

20
00:04:43.330 --> 00:04:56.300
Some discussed societal impacts include the facilitation of filter bubbles or echo chambers. So there is groups on social media platforms that are increasingly disconnected from others and hold polarized views on societal issues.

21
00:04:56.520 --> 00:05:03.840
This year has been criticized as a technologically but deterministic way of algorithmic functioning. And there has been little evidence for this.

22
00:05:03.850 --> 00:05:10.939
Rather algorithms are understood to reflect and reproduce polarizations and inequalities that already exist in the offline world

23
00:05:12.070 --> 00:05:13.390
and

24
00:05:14.160 --> 00:05:29.899
the structure agency continuum as discussed by Linda, presents another way to look at algorithms. And this viewpoint re-emphasizes human agency and algorithmic functioning. And Lyndal applies audio's theorizations to investigate the social power of algorithms.

25
00:05:29.960 --> 00:05:44.100
And the idea here is that social media platforms hold algorithmic meta-capital that is platform sold to centralized power over the visibility of the content, and thereby subtly shaping the habit of its users in the case of seeking visibility.

26
00:05:44.210 --> 00:05:55.359
So there, yeah, seeking visibility to and through the algorithm. And sadly, as The normative rules are not explicit, and they are also constantly changing.

27
00:05:55.760 --> 00:06:06.309
And following this view, users need to rely on their gut feeling for the game and the algorithm sky uses towards certain actions and but they do not force them to gain symbolic capital

28
00:06:09.410 --> 00:06:29.400
and these notions have us to understand how political information is shared by social media influences. And in my research I adopted a qualitative research design. And there are 2 things I want to briefly highlight here in the lead up to discussing some of my findings. So on the right side, you see some profiles of my participants that I interviewed.

29
00:06:29.490 --> 00:06:32.770
and what I wanted to highlight here is that I used a case study

30
00:06:32.780 --> 00:06:43.329
as a contextual framework to identify political content online. And I selected the repeal of paragraph 219, A of the German Criminal Code as a case study.

31
00:06:43.330 --> 00:07:10.699
And this is a sub paragraph in the abortion legislation in Germany, and it prohibited medical professionals from publicly informing or advertising whether they perform abortions and the methods they used, and the paragraph has been repealed in June last year by the German Parliament. And the main take way here is that the repeal presents an improvement in from information accessibility, and it is seen as a step towards a total decriminalization of abortions in Germany.

32
00:07:10.940 --> 00:07:21.960
and it's everywhere abortion rides a highly contested issue, and the debate drew attention from a wider society, including also the influences who shared their opinions regarding the debate.

33
00:07:22.850 --> 00:07:35.210
Secondly, I applied an interview technique called trace interviewing which uses the visualizations of participants, digital traces, for example, their posts to initiate reflections on their content, and how they presented it.

34
00:07:37.170 --> 00:07:52.310
So far I've conducted 7 interviews with influencers who publicly supported the repeal of the abortion law. My participants were between 30 to 60 years old, and can be classified as Nano and micro influences with audiences between just over a thousand to over 12,000 followers.

35
00:07:53.070 --> 00:08:01.710
So presumably my interviews are part of a quite large number, maybe of active users to which the algorithm might not be as responsive to when considering the numbers of followers.

36
00:08:02.210 --> 00:08:12.140
and my participants are interested in current social, political issues with some folks on current political news and others. Some are interested in feminism, finding climate change, and so on.

37
00:08:12.230 --> 00:08:18.799
some primarily share political content, and others share a mix between political and other content.

38
00:08:19.620 --> 00:08:32.059
 yeah. And now I will share and discuss with you 2 themes that emerged in my preliminary analysis of the interviews, exploring the interplay between influences and the algorithmic visibility.

39
00:08:34.780 --> 00:08:48.260
So me and the platform universe tensions, navigating and sharing political content covers influences conceptualizations of their online profile, entangled in the affordances and in between other contents of social media platforms.

40
00:08:49.060 --> 00:09:16.330
from an influences perspective whose sharing and also consuming political and other content online. There are experienced tensions between, for example, finding reliable information and encountering fake news feeling as part of community and being subject to online hate, experiencing platforms as an accessible and easy way to share political work and opinions, but also for some, the only way to share such work and opinions, making them feel dependent on platforms.

41
00:09:16.460 --> 00:09:22.809
which goes along with influences and dying, using the platforms as it's not in trying using it.

42
00:09:23.050 --> 00:09:30.719
and influence us to capitalize on similarity, but also some who seek dissimilarity or dissimilar opinions online.

43
00:09:30.790 --> 00:09:48.940
and there are careful placing of political content publicly while being critical about the economies of platforms at the same time. And most importantly, here. they have a sense of agency, but as as well as powerlessness when they navigate the space online.

44
00:09:49.260 --> 00:10:15.339
so very broadly. These tensions highlight the power held by platforms, and particularly looking at the negotiating visibility, the tension agency versus powerlessness, the the bottom. One supports this idea of the algorithmic meta-capital head by platforms which gives influencers guessing and adapting their behaviors to the constantly changing rules of visibility and social media platforms.

45
00:10:16.710 --> 00:10:23.350
and the next theme we take a closer look at different ways, and Frances express the agency when negotiating the content

46
00:10:26.180 --> 00:10:29.920
and the quote up here inspired the themes title and goes

47
00:10:29.930 --> 00:10:45.730
so it's not like you in a downtown area, and you go into some store and check it out and then continues, but I feel like they're explicitly going somewhere because they want to go there figuratively. And this is by one of my interviews, Lucy Langston, describing how her followers find her account.

48
00:10:45.760 --> 00:10:52.140
and she also continued saying. If I want to get more followers again, I'd have to adapt.

49
00:10:52.920 --> 00:11:22.280
and I've put up some quotes here. It's a lot but that illustrate different interpretations and ways. My interviews expressed agency when they shared political content on Instagram, and we'll go through them and group them now. so the first 2 display wall is strategically using hashtags. to position one's content exemplified here by the quote by, Hey, we can do it using hashtag bombing.

50
00:11:22.280 --> 00:11:36.100
and what she describes as hashtag bombing, which is using the maximum number of hashtags that are available on Instagram and my other participant using hashtags to break into groups with dissimilar opinions.

51
00:11:38.840 --> 00:11:58.780
the next strategy was to exploit the affordances and options available on the platforms. first, of like tailoring the content to the required format, and also making use of the different ways to share, to share content, including using posts and stories, story highlights, PIN stories, and so on.

52
00:12:00.520 --> 00:12:07.980
then most of my participants timed their posts so much times where they see most people use the platforms, for example, during commuting times.

53
00:12:08.410 --> 00:12:21.530
And then some of my participants are also distance themselves from the platforms in order to for them to create the best content possible without consciously thinking of what the algorithm might prefer

54
00:12:22.730 --> 00:12:24.380
and oops.

55
00:12:24.900 --> 00:12:45.370
Lastly, some of my interviews heavily drew on their social capital. So they online as well as offline networks to gain visibility and reciprocate visibility. And this reciprocity of visibility was, I thought, really interesting is my interviews kind of leverage their social capital to gain more like visibility or symbolic capital online.

56
00:12:45.590 --> 00:13:00.710
And one of my participants also said, It's more like a give and take So she. She was quite active in liking and commenting other posts, and thereby had the feeling of she would gain more visibility. for her content as well.

57
00:13:01.760 --> 00:13:11.189
so influences consciously make decisions and adapt when sharing political content online in response to the assumed rules of visibility of social media platforms.

58
00:13:11.250 --> 00:13:24.219
And, to put it in bodious terms, it appears that platforms shape the habit of its users. In this case the influencers who consciously apply different strategies, adapting their behavior and trying to play the visibility game on social media platforms.

59
00:13:25.130 --> 00:13:26.290
And

60
00:13:27.250 --> 00:13:41.180
in conclusion, the idea of human agency and algorithmic functioning and applying video's ideas on meta-capital to algorithmic visibility provides a valuable new way of understanding and exploring the interplay of humans and algorithms on social media platforms.

61
00:13:41.480 --> 00:13:53.720
the discussed tensions for navigating and sharing political content on Instagram, particularly regarding, gaining and maintaining disabilities, support this idea of Platform's algorithmic manor-capital essex.by lunar

62
00:13:53.850 --> 00:14:04.110
and by applying a range of strategies to play the algorithm influences expressed the sense of agency when they navigated visibility within the bounds afforded by social media platforms

63
00:14:05.620 --> 00:14:29.959
and further research, I think, is needed to investigate how they are also constantly changing platform environments and the development of habiters interact over time as it is a very fast-paced environment. And another, it might be interesting aspect, was the last point on the previous slide, the formation of online networks and employing kind of the social capital to gain visibility, which is something I want to explore in later analysis.

64
00:14:30.250 --> 00:14:32.110
Yeah, thank you.

65
00:14:37.480 --> 00:14:39.180
For a time management.

66
00:14:41.150 --> 00:14:49.529
I didn't say that upfront. But let's continue with all the presentations, like in other sessions, and then and end with a panel.

67
00:14:50.240 --> 00:14:53.600
So now this the real deal, because

68
00:14:53.740 --> 00:14:59.979
I've been doing a zoom dance hall the whole afternoon, not to mix up with the Tik Tok dance.

69
00:14:59.990 --> 00:15:03.680
as in the hopes that this is going to work

70
00:15:03.990 --> 00:15:10.710
for the ones who were here this morning. let me first see how I need to stop share.

71
00:15:11.180 --> 00:15:13.780
And then. yeah.

72
00:15:15.340 --> 00:15:21.929
and remote presenter can share this screen. Yeah, yeah, they already tried. So it should work. But we had a lot of

73
00:15:22.570 --> 00:15:24.529
sound issues. But

74
00:15:24.540 --> 00:15:29.650
I see you do this, Milda. So if you could share your screen

75
00:15:30.710 --> 00:15:34.500
and unmute, of course. Done.

76
00:15:41.850 --> 00:15:46.929
Dorismilda Flores-Márquez: Okay, yes. Are you seeing my screen? Yes.

77
00:15:48.130 --> 00:15:57.330
Dorismilda Flores-Márquez: Great. Well, Hello, everyone. I'm Doris Mila, Florence Marquez, Ramon, your Sierra de La Saliva here in Mexico.

78
00:15:57.390 --> 00:16:09.879
Dorismilda Flores-Márquez: and let me start with a confession or more. First is almost 2 in the morning here in Mexico, so if I not sound so smart. I will blame the night, of course.

79
00:16:09.970 --> 00:16:32.789
Dorismilda Flores-Márquez: a second in terms of positionality as researcher. I'm a Latin American communication scholar. My educational background is in communication and sociology. And my research interests are detail, culture and communication activism, social change in terms of the ital public expression of activists.

80
00:16:32.950 --> 00:16:47.349
Dorismilda Flores-Márquez: This is important for what I am going to say later, and the your confession is The last 7 years of my life a. I have been in a kind of an infinite loop of projects.

81
00:16:47.600 --> 00:17:03.560
Dorismilda Flores-Márquez: a do to university dynamics, you know. I finish one project and I start another one. And this is not a complaint. Of course I think I have been very lucky to have funding for research, but sometimes it is also exhausting.

82
00:17:03.620 --> 00:17:08.850
Dorismilda Flores-Márquez: But just when the conditions at my University change it.

83
00:17:09.010 --> 00:17:27.190
Dorismilda Flores-Márquez: and we have a little more time to work on our research projects and a little more freedom as well. A kind of explosion of Ivs. It started. So what I am going to present comes from these ideas. I'm from my permanent, interesting metalogue

84
00:17:27.710 --> 00:17:29.330
Dorismilda Flores-Márquez: in this way.

85
00:17:30.970 --> 00:17:37.729
Dorismilda Flores-Márquez:  the presentation is not based on certain things, but on questions.

86
00:17:37.770 --> 00:17:47.249
Dorismilda Flores-Márquez: Those who research a detailed practices know that we tend to focus on the presence of actors in detail platforms and their detailed practices.

87
00:17:47.410 --> 00:17:53.399
Dorismilda Flores-Márquez: This implies at least 2 aspects. First, we see what is visible.

88
00:17:53.540 --> 00:17:58.159
Dorismilda Flores-Márquez: and second, we focus mainly on those sectors who have or have had

89
00:17:58.210 --> 00:18:16.969
Dorismilda Flores-Márquez: they will to be in a detailed space in a digital platform doing digital practices. But what about those actors that appear online without having decided that? And what about those actions that leave traces even without the actors being aware of it?

90
00:18:17.580 --> 00:18:38.150
Dorismilda Flores-Márquez: So in in times of him, thatification, platformization, algorithmization, it seems almost impossible for anyone to be out of the world of digital data. So there is a gap about a absence and a how to approach it. In times of algorithms.

91
00:18:38.320 --> 00:18:50.480
Dorismilda Flores-Márquez: I have been researching the presence and particularly the public presence and the public expression for years. But I am kind of a novice exploring the absence.

92
00:18:50.600 --> 00:18:57.449
Dorismilda Flores-Márquez: So I explore 2 cases of absence that even so in have left the the traces.

93
00:18:57.710 --> 00:19:20.179
Dorismilda Flores-Márquez: In these cases, let me think on the implications of being there or not, in order to generate ideas about how to study absence. And this is the perfect opportunity to put the a mythological imagination into action, because it's relatively easy to approach presence, but how to? As to the absences?

94
00:19:20.630 --> 00:19:25.859
Dorismilda Flores-Márquez: let me start with a sociological imagination.

95
00:19:26.060 --> 00:19:46.250
Dorismilda Flores-Márquez: According to right meals. A the sociological imagination is the possibility of connecting the biography with the history, understanding the interplay between them and contextualizing personal experiences into brother instructors. In other words, this means avoiding

96
00:19:46.340 --> 00:19:52.860
Dorismilda Flores-Márquez: simplistic explanations in order to look for situated and contextualized it. And explanations

97
00:19:53.090 --> 00:20:01.390
Dorismilda Flores-Márquez: a. This has, of course, mythological implications, since it requires not to take anything for granted.

98
00:20:01.730 --> 00:20:09.869
Dorismilda Flores-Márquez: But to think of methodological options that reflect the complex relation between actors, practices, and context.

99
00:20:10.050 --> 00:20:15.739
Dorismilda Flores-Márquez: In this line methodological imagination implies going beyond

100
00:20:15.800 --> 00:20:27.500
Dorismilda Flores-Márquez: the classical mythological tools to transform my that or create new resources or techniques that let us approach the complexity of a our object. So if I study.

101
00:20:28.470 --> 00:20:39.150
Dorismilda Flores-Márquez: and in this time I have found a inspiration in some texts. so I share here some methodological notes about it.

102
00:20:39.220 --> 00:20:52.120
Dorismilda Flores-Márquez: First, Andrew Abbott, who said that science is a conversation between rigor and imagination. as provided a number of reflections and methodological options to produce new knowledge.

103
00:20:52.270 --> 00:21:00.360
Dorismilda Flores-Márquez: So I am going to briefly take up a 2 ideas from a his book, methods of discovery.

104
00:21:00.450 --> 00:21:10.589
Dorismilda Flores-Márquez: that is, a first ideas, according to about a most ideas, come from looking at data.

105
00:21:10.900 --> 00:21:16.660
Dorismilda Flores-Márquez: And once you've got an idea, you need to try it out on some new date.

106
00:21:16.930 --> 00:21:18.570
Dorismilda Flores-Márquez: So,

107
00:21:19.730 --> 00:21:24.939
Dorismilda Flores-Márquez: this is a kind of a circle between the

108
00:21:25.040 --> 00:21:27.559
Dorismilda Flores-Márquez: a test of these ideas

109
00:21:27.630 --> 00:21:30.610
Dorismilda Flores-Márquez: and the bus list

110
00:21:30.690 --> 00:21:38.709
Dorismilda Flores-Márquez: according to to about as well a research when it is done with rigor and imagination, it's like a postly.

111
00:21:38.850 --> 00:21:46.060
Dorismilda Flores-Márquez: because we have empirical, positive pieces, and we work with them to complete them, or to find a solution

112
00:21:46.770 --> 00:21:48.500
Dorismilda Flores-Márquez: and

113
00:21:48.530 --> 00:21:56.019
Dorismilda Flores-Márquez: other a text that was in kind of inspiring for this, Explorations? Is this one a.

114
00:21:56.370 --> 00:22:01.050
Dorismilda Flores-Márquez: And it's got

115
00:22:01.140 --> 00:22:02.530
Dorismilda Flores-Márquez: and

116
00:22:03.140 --> 00:22:09.350
Dorismilda Flores-Márquez: he proposes 10 meteorological considerations. A developed council analysis

117
00:22:09.380 --> 00:22:21.749
Dorismilda Flores-Márquez: extend and challenge explanatory symmetry specify the type of form of absence under investigation, define the subject, identity, face case parameters make sure the shapes of absence, make sure relations of absences.

118
00:22:21.760 --> 00:22:24.230
Dorismilda Flores-Márquez: make sure densities of absences

119
00:22:24.300 --> 00:22:34.070
Dorismilda Flores-Márquez: that you comparison, and you to remove. This awareness is kind of a different, but it is very important to think of

120
00:22:34.490 --> 00:22:36.980
Dorismilda Flores-Márquez: how to approach a absences.

121
00:22:37.300 --> 00:22:46.089
Dorismilda Flores-Márquez: And now let me present a 2 cases to ideas and pieces of data that are parts of a postly.

122
00:22:46.190 --> 00:23:00.429
Dorismilda Flores-Márquez: and that are different forms of absence, just as a kind of disclaimer. There is no rigorous data collection or data production here. These are just 2 cases that let me think about a absence.

123
00:23:00.750 --> 00:23:15.730
Dorismilda Flores-Márquez: And This is the case when this is a guy, We can call him the one who's nothing he decided not to have social media profiles because he doesn't like to expose his privacy

124
00:23:15.790 --> 00:23:27.890
Dorismilda Flores-Márquez: and his way of protecting it. If you stay out of the way. however, I found traces of him accidentally, because Facebook suggested me a guy with his name in people, you may know.

125
00:23:28.260 --> 00:23:31.219
Dorismilda Flores-Márquez: it was not him, but his father.

126
00:23:31.280 --> 00:23:33.299
Dorismilda Flores-Márquez: A, who has the same name.

127
00:23:33.320 --> 00:23:44.009
Dorismilda Flores-Márquez: but a few clicks allowed me to find a photos of him uploaded by his his father, where his wife, who does have social media profiles is stacked as well.

128
00:23:44.140 --> 00:23:58.709
Dorismilda Flores-Márquez: and with a few more clicks it was possible to find their and their wedding photos. So they one who has not been him is there, and the privacy that he protects by

129
00:23:58.890 --> 00:24:10.060
Dorismilda Flores-Márquez: not being in social media is just an illusion, because even even a very private moment like his wedding is exposed online beyond his will.

130
00:24:10.810 --> 00:24:18.080
Dorismilda Flores-Márquez: A. This case shows a a voluntary absence that in practice is an involuntary presence.

131
00:24:18.220 --> 00:24:33.790
Dorismilda Flores-Márquez: And this case, let us reflect on right. It's such as privacy direct to protections of one c mash. But let us also reflect on thatification and the pervasiveness of the Ital platforms in social life.

132
00:24:34.340 --> 00:24:38.000
Dorismilda Flores-Márquez: let me show now the case to.

133
00:24:38.090 --> 00:24:41.299
Dorismilda Flores-Márquez: as you may see. This is a map with the route.

134
00:24:41.400 --> 00:24:52.439
Dorismilda Flores-Márquez: My team and I went on a field trip to the Northern Sierra of in Mexico. We were doing field work for a study about a digital inequalities

135
00:24:52.730 --> 00:24:54.210
Dorismilda Flores-Márquez: in rural areas.

136
00:24:54.230 --> 00:24:58.680
Dorismilda Flores-Márquez: and in this area there is no mobile phone.

137
00:24:58.820 --> 00:25:02.439
Dorismilda Flores-Márquez: So no matter how many precautions we took, we

138
00:25:02.450 --> 00:25:06.319
Dorismilda Flores-Márquez: the that anyone out with the we got lost

139
00:25:06.570 --> 00:25:07.880
Dorismilda Flores-Márquez: and

140
00:25:08.130 --> 00:25:23.470
Dorismilda Flores-Márquez: as GPS works beyond the mobile signal. The route we followed kept registered in Google Maps and we Retr. They say that hours later. This is a very different case, but we left traces, even without intending to.

141
00:25:23.750 --> 00:25:27.059
Dorismilda Flores-Márquez: This case, shows us that kind of

142
00:25:27.069 --> 00:25:35.789
Dorismilda Flores-Márquez: kind of pheno centre. P. Set again a presence where digitally registered, even when we were not intending to

143
00:25:35.870 --> 00:25:48.580
Dorismilda Flores-Márquez: when we were not aware of it. And in this case a let us reflect on thatification that the collection, privacy, and the pervasiveness of the the platforms in social life as well

144
00:25:49.330 --> 00:26:13.759
Dorismilda Flores-Márquez: and in bad cases. A first, they have the data and the intuition, and later, I think that with a theoretical explorations, and let me talk first about presence and absence, I said before that I have been analyzing the public expression practices involving the presence of factors in the Italian networks and beyond several years.

145
00:26:13.810 --> 00:26:21.919
Dorismilda Flores-Márquez: And these activists and the activists I have worked with in this type, whether they want to be present

146
00:26:22.149 --> 00:26:26.880
Dorismilda Flores-Márquez: in networks and express themselves publicly or not

147
00:26:27.100 --> 00:26:30.629
Dorismilda Flores-Márquez: in terms of a

148
00:26:30.840 --> 00:26:38.560
Dorismilda Flores-Márquez: The production of presence recognizes the materiality in human experience. So in the practices.

149
00:26:38.680 --> 00:26:39.980
Dorismilda Flores-Márquez: this is some

150
00:26:40.060 --> 00:26:51.939
Dorismilda Flores-Márquez: a digital production of presence being there being online being connected is a digital production of presence. But again, what about absence.

151
00:26:52.010 --> 00:27:05.179
Dorismilda Flores-Márquez: Considering that 2 cases presented, there is a digital construction in of absence as well. That is not really an absence, because data is still being produced.

152
00:27:05.410 --> 00:27:10.409
Dorismilda Flores-Márquez: Whether we want it or not, whether we know it or not.

153
00:27:10.520 --> 00:27:18.889
Dorismilda Flores-Márquez: and that so not being there, not being voluntary online not being connected also processes, data

154
00:27:19.190 --> 00:27:28.379
Dorismilda Flores-Márquez: either from others, people's day, that on from its own data in the production in the background. and.

155
00:27:28.750 --> 00:27:38.050
Dorismilda Flores-Márquez: on the other hand, a proposed to think of technologies of life in the Spanish. He say, a technologies

156
00:27:38.260 --> 00:27:42.819
Dorismilda Flores-Márquez: to refer to the mentality of these technologies in everyday life.

157
00:27:42.960 --> 00:27:53.729
Dorismilda Flores-Márquez: It. This progressiveness of this data technologies implies that we no longer connect to them as a state between offline and online awards.

158
00:27:53.770 --> 00:28:01.880
Dorismilda Flores-Márquez: but that we are always connected and received notifications. For example, as a stimulus to maintain those connections all the time.

159
00:28:02.070 --> 00:28:08.389
Dorismilda Flores-Márquez: and so being there is thing there almost permanently or permanently.

160
00:28:08.760 --> 00:28:09.920
Dorismilda Flores-Márquez: And

161
00:28:10.260 --> 00:28:20.710
Dorismilda Flores-Márquez: to conclude a while, this is still to be studied in greater detail where I have been explored and reflected so far.

162
00:28:20.720 --> 00:28:34.000
Dorismilda Flores-Márquez: allow me to say that in times of algorithms this pervasiveness of people. Technology in everyday life is switch that absence. A can be read as a different, a form of presence.

163
00:28:34.110 --> 00:28:39.600
Dorismilda Flores-Márquez: Things there being online being connected is a digital production of.

164
00:28:39.800 --> 00:28:46.570
Dorismilda Flores-Márquez: but not being, there is a digital production of.

165
00:28:46.600 --> 00:28:56.999
Dorismilda Flores-Márquez: So in the life, there is no practical position between presence and absence, because real absence. Not it does not exist

166
00:28:57.070 --> 00:29:03.329
Dorismilda Flores-Márquez: any more. So thank you so much for your attention.

167
00:29:09.950 --> 00:29:30.020
Thank you very much for the wonderful presentation, and I hope, even though it's a a terrible time at your place. You can stick around for the for the discussion later. Yeah, 6 coffees. And then I would like to invite justice to share his video.

168
00:29:30.780 --> 00:29:35.080
We have all options here today, videos, everything.

169
00:29:39.350 --> 00:29:42.240
But can you see my screen guide, though

170
00:29:42.700 --> 00:30:08.400
Justus Rahn: perfect. So without further, I do, I would start the video as we have a pre recorded session, due to the fact that we were actually worried that we would have some technical issues during the live presentation. And because, both Katarina and Kobinia cannot make it in person today. So let's just start the video. And once you have question we'll discuss it later in the now

171
00:30:10.840 --> 00:30:22.339
Hello, and a warm welcome from Hanover, Germany to all Isa participants. We are excited to share our insights on discuss strategies for the implementation of algorithmic decision support systems. Today.

172
00:30:22.570 --> 00:30:33.099
we have 3 qualitative-oriented sociologists from Lyman's University in Hanover, exploring digitalization as a heterogeneous process in public organizations.

173
00:30:33.110 --> 00:30:48.329
Yocto, Sr. Who is currently speaking Kobiniangara and Qatar Bounceman. We want to show how the Austrian public employment service succeeds in not only technically implementing an algorithmic support system, but also discursively manage that implementation

174
00:30:48.370 --> 00:31:02.529
facing a public that is very, very skeptical towards algorithmic decision support, especially if, as in our case, an unemployed person is categorized for his or her eligibility for welfare funding upon his demographic input

175
00:31:03.200 --> 00:31:14.250
within the next 50 min we would like to give a presentation based on a paper of the same name that was published in a special issue on digital transformation of historical social research. In 2022,

176
00:31:14.640 --> 00:31:30.450
we will firstly situate our research in streams of algorithms and organizations on the one hand, and technology implementation in public discourse. On the other hand. Afterwards we will plausible our research program, sociology of knowledge approach to discourse.

177
00:31:30.560 --> 00:31:49.089
and we used it for the analysis of all media coverage through which the public discourse was constructed. Presenting our findings, we identified 3 major disco strategies that helped the organization. Am. S. Scanning public legitimacy and immunizing the algorithm from skeptical or critical voices over time.

178
00:31:49.520 --> 00:31:56.740
Finally, in the conclusion, we reflect the findings in the light of public discourse and explain the immunization of the algorithm.

179
00:31:57.780 --> 00:32:07.800
As I pointed out in the introduction. There are 2 research streams guiding our thoughts. Both acknowledged the intertwined relationship of the panel's title Algorithm and society.

180
00:32:07.940 --> 00:32:24.160
and we want to unfold this acknowledgment by following Gsp. Stating that algorithms are never as objective as they seem. Instead, they are often created by heterogeneous teams that sometimes have diverging interests and that are working under certain regulatory constraints.

181
00:32:24.840 --> 00:32:37.520
Algorithms are for most developed, tested, and deployed in organizations. They operate in around with or through organizations, and without them these technologies could never achieve their social impacts.

182
00:32:37.820 --> 00:33:00.019
This is shown by Buchname Dostyle will understand algorithms as digital observational formats. Organizations can act by embedding such technologies at sensitive, socio-political, charged. But decision points in public administrations, and we learn from those works to look closely at organizations as the place where technologies come to life.

183
00:33:01.080 --> 00:33:14.049
The second one deals with the discursive aspect of technology innovation. It is often recognized as a solution for social problems. But to surface a solution that need to be discursively prepared to be accepted.

184
00:33:14.660 --> 00:33:27.150
Akber argues that the success of German predictive policing applications can be explained, not least by the fact that residential burglaries have been publicly unfolded as a large-scale security problem.

185
00:33:27.400 --> 00:33:42.070
And the solution for that large scale security problem is the pre-crime algorithm having said that we understand technology as a discursive phenomenon that requires special attention due to its situatedness in social administration organizations

186
00:33:43.320 --> 00:33:53.659
bringing these 2 threats together, we ask, what role does the Ims organization play in shaping the discourse of implementation of its own algorithmic decision support system

187
00:33:55.480 --> 00:34:10.530
to analyze that we use sociology of knowledge approach to discourse, in short, Scots that combines the social constructivism of Baghaludman with discourse and analysis of Michelle Focour, and is to be understood as a research program rather than a rigid methodology.

188
00:34:10.610 --> 00:34:16.739
Knowledge in that sense is shaped through discourse, through collectively shared structures of interpretation.

189
00:34:17.890 --> 00:34:29.369
We opt for scar. For 2 reasons it focuses on the social situations of digital technologies. meaning the nature of the algorithm is negotiated by the dispos carriers.

190
00:34:29.500 --> 00:34:43.500
and whoever accumulates the most power of definition excludes realities other than its own. The second reason is that it allows a processional view. Scout takes changes and shift in the discourse into account.

191
00:34:44.630 --> 00:34:54.919
breaking it down. We analyze the data Corpus, including 170 documents, most of the newspaper articles who focus on the implementation of Ams algorithm.

192
00:34:55.260 --> 00:35:12.820
They ranged from 2,018 to 2,020, and by doing so we identified the strategic part of knowledge politics. That means the role of processes and actors with their interests and strategies that go through, stabilize, and change the structure.

193
00:35:13.290 --> 00:35:20.860
In short, these can be described as discourse strategies that expand the power of definition for their respective actors and discourse.

194
00:35:21.990 --> 00:35:30.979
Now, Katar Onceman will present our findings. 3 discourse strategies that shape the discourse on the implementation of the Ams Algorithm.

195
00:35:32.080 --> 00:35:52.329
Thanks yoursters. At first some words about our research case, the I Ms algorithm, the algorithm was developed to predict labour market opportunities as a region-specific chances for labour market integration and classify unemployment persons with logistic regions.

196
00:35:52.330 --> 00:36:04.980
Our empirical case relates to the discourse around the introduction of an algorithmic decision support system in the labor market service in Austria. Short Ams

197
00:36:05.370 --> 00:36:15.670
like you just had, between 2018 and 2020 an intense public discourse about the Ams. And the algorithmic decision support system developed.

198
00:36:15.880 --> 00:36:26.410
It is characterized by different critiques, topics, group of actors, and speakers with different degrees of power. We come back back to this and our findings.

199
00:36:26.460 --> 00:36:38.969
The overview of the discourse shows a strong clarity as a fundamental dynamic. It consists of the partials from the algorithm who are close to the Ms.

200
00:36:39.170 --> 00:36:54.420
And critics who are academics, journalists, or members of Ngos. In the following we show 3 discourse strategies that Ms. As an organization has used to configure the discourse around its own algorithm.

201
00:36:54.620 --> 00:37:05.500
It starts with the first strategy which we called place, which we call placing algorithm and the discourse of field of welfare, state support and efficiency.

202
00:37:05.980 --> 00:37:25.509
We reconstructed that welfare, state apart, is used as a defuse reference point for various augmentations. and these diffuse references show up, firstly, along organizational internal expectations of the Ms regarding the algorithm.

203
00:37:25.820 --> 00:37:38.590
Thus the algorithm is framed by the support us as an assistant system of helping it is used to help classify employment persons in the chancelling process.

204
00:37:38.980 --> 00:37:44.009
The algorithm is located in the organization as a part of it.

205
00:37:45.480 --> 00:37:57.740
and on the other side the defuse reference are, secondly, shown along external expectations which refer to the organization as a whole, not on the algorithm itself.

206
00:37:57.940 --> 00:38:08.919
Critical discourse positions expect the counseling to improve the individual situation of the unemployment person. At that time they don't

207
00:38:09.300 --> 00:38:22.619
develop their own interpretation of how the algorithm itself can have clients. Moreover, we observe that the algorithm is placed in a discourse with the expectation

208
00:38:22.640 --> 00:38:32.840
of these dynamics will fundamentally shape the discourse. They are also part of the second is closed strategy which I now describe

209
00:38:33.350 --> 00:38:47.650
as a second strategy. We reconstructured referential agility in the conception of the algorithm. the so-called S algorithm, and I go several reconfigurations. During this course

210
00:38:48.170 --> 00:39:08.829
the Foo's description as a computer, a mathematical model, or a big data make the algorithm a moving moving target in the discourse. And at the beginning the algorithm was introduced as a complex tool that works by means of highly highly complex mathematical models

211
00:39:09.020 --> 00:39:15.840
in the same as cause fragment. More imagines of the algorithm are introduced. Now I quote.

212
00:39:17.130 --> 00:39:27.130
recognizing specific patterns and so cold big data in order to make predictions from them is a skill that computers are clearly better at

213
00:39:27.320 --> 00:39:38.299
than humans. The concept of algorithm is unclear because of the conflicting tape terms. And it is it's days that rehearsal over the whole discourse

214
00:39:38.960 --> 00:39:57.989
the remarkable range of argumentative imagines and settings clue to Kurdish course positions. This dynamic makes the algorithm the algorithm's implementation more defensible and it is protected against critical positions over a long period of time.

215
00:39:58.030 --> 00:40:01.429
The last strategy is now presented by Kobinian guy.

216
00:40:03.040 --> 00:40:16.100
Thanks, Cutter. This third strategy leads to the fact that even the critics in the Ms. Discourse abandoned their criticism of individual functions and programming of the algorithm.

217
00:40:16.810 --> 00:40:32.539
The reason, therefore, lies in a shift in criticism which now referred very broadly to the as an organization and its processes. The algorithm as a target for criticism faded into the background of the Ms. Discourse

218
00:40:32.960 --> 00:40:42.379
to put it bluntly. the algorithm was stabilized as a part of the organization, and there were no longer any attempts to diligit demise.

219
00:40:43.430 --> 00:41:00.240
We will show you this with an example. With the controversy surrounding the algorithm. Growing early in the Ems discourse. there were indications that it was being discursively shifted into the organization in November 2,018,

220
00:41:00.430 --> 00:41:16.340
a spokesperson of the Then said. Please be assured it is still the person. the ums counselor who personally takes care of the concerns of the job seeker. But who can now help even better?

221
00:41:17.540 --> 00:41:28.710
This statement became a key event of the discourse as the Ems Board began to frame the algorithm as an organizational technology. And we will see why?

222
00:41:28.760 --> 00:41:31.290
Because the discourse jumps on it.

223
00:41:33.020 --> 00:41:44.989
I quote what information and the grease of freedom does the as Councillor have to disregard the machine's decision in example to make the judgment himself.

224
00:41:46.270 --> 00:42:02.159
This asks a computer science professor influential in the discourse in a daily newspaper thems thus becomes the address in a double sense, not only of programming also, in applying the algorithm.

225
00:42:04.580 --> 00:42:21.059
I summarize by discussing where, how and whom the assistant system helps in that terminates countering in the understanding of what the algorithm actually is, inscribed itself into the Ems discourse

226
00:42:21.930 --> 00:42:39.560
game. S. Board hat uses this referential agility. To avoid criticism of, for example, the distribution of welfare, state assistance, or demands for transparency, and even expands this uncertainty about the subject of the discourse.

227
00:42:40.770 --> 00:42:51.579
The third discourse strategy described is incorporating the algorithm by the The algorithm is thus protected from criticism.

228
00:42:51.800 --> 00:42:57.960
Instead. the critique is framed within the context of the organization and its processes.

229
00:42:59.110 --> 00:43:18.679
These discourse strategies be a fruit as attempts at the legitimization decrease over time previously hoisted into the public eye as a prestigious project, then hemp back into the organization. The algorithm becomes harder to address as an object of public criticism.

230
00:43:19.330 --> 00:43:31.079
drawing a conclusion despite its importance. The use of the algorithm in the organization's context remains in its invisible sphere of decision.

231
00:43:31.830 --> 00:43:43.370
For this reason we encourage further scholarly efforts to gain insight into implementation processes around all algorithms in public organizations.

232
00:43:45.480 --> 00:43:55.950
At the end, we want to thank the organizers of this panel for creating a platform where we can discuss the implications of algorithms on our society and vice versa

233
00:43:56.640 --> 00:44:11.430
as well. We want to express our gratitude to everyone listening. If you want to ask questions about our paper. which is open access under the presented link. You can reach us under the displayed contact details.

234
00:44:12.200 --> 00:44:14.750
We sent one greetings to Melbourne.

235
00:44:25.890 --> 00:44:30.580
so that worked very well. Thank you for doing that in great presentation.

236
00:44:30.800 --> 00:44:47.290
Both online presentation survived. What a success! And now we go back to just a quick remark. If you want to, I would place the the of the respective favor and to the public changes that fine with you.

237
00:44:47.830 --> 00:44:56.420
Yeah, of course. Yeah. So the people in line can for sure see it? and otherwise they can get in touch with you.

238
00:44:58.180 --> 00:45:10.160
 okay.

239
00:45:27.990 --> 00:45:31.750
So I need to be in view of the camera. Right? As I thought, yeah, okay, so I'll

240
00:45:32.110 --> 00:45:55.009
that's okay. No, I'm not that ambitious. But I just want to be able to see. And so well, good evening, everybody. On Friday, the almost last day of the conference, and I'd like to start by thanking you for coming, and to Atila and Lauren for for being here and coordinating everything, and and making this possible. And

241
00:45:55.330 --> 00:46:02.110
so here we go. My name is Eric Schneiderh. I'm a professor at the University of Toronto in the Sociology department.

242
00:46:02.170 --> 00:46:05.809
and the research I'd like to talk to you about today

243
00:46:07.700 --> 00:46:13.019
is part of a book project. Now, this is not advancing. Do I do it with the mouse, or do I?

244
00:46:13.230 --> 00:46:32.659
There we go. Yeah, that works. It's part of a book project with my co-author, Martin look who could not be here because he needed to go back to Canada for a wedding. So he left early this morning. and it's coming out in September with Stanford University Press and the book looks at

245
00:46:32.850 --> 00:46:44.129
how people use the Internet. And specifically, the website, Gofundme Com. to raise money for costs like medical bills and tuition and emergency expenses.

246
00:46:44.350 --> 00:46:45.970
In the book we answer a

247
00:46:46.010 --> 00:46:51.390
number of questions about this kind of fundraising, including where it came from, who uses it

248
00:46:51.430 --> 00:46:59.930
and what the process is like, and and whether people get the help they need, and the title tells you that they they tend to not get the help they need.

249
00:47:00.210 --> 00:47:16.229
One of the topics we emphasize is the important role of algorithms and what they play in the process. And this is the focus of today's presentation. I'm going to start by briefly discussing some key areas where scholars have documented the use of algorithms and then introduce our case

250
00:47:16.280 --> 00:47:32.699
crowdfunding and show out fits into this discussion. Then I'll discuss our analyses and what we found before closing with some broader implications, and then I have a little add on, because my talk is perfectly timed. I found some connections with all the 3 other papers which it was I'm really excited about. Actually

251
00:47:35.070 --> 00:47:43.060
so. scholars have documented the use of big data and algorithms which most of you know, are forms of automated decision making

252
00:47:43.190 --> 00:47:56.579
in many different parts of the web, including in search journalism and gig work, in ways that promote problematic stereotypes, enable oppressive labor practices and even spread political disinformation.

253
00:48:00.750 --> 00:48:27.980
While many of these studies have looked at how algorithms are used by large technology companies like Uber or Google. scholars have increasingly drawn attention to how algorithms are used in the public sector in similarly problematic ways, government agencies like unemployment offices. It's similar to what we were just hearing about. I use algorithms to determine who to audit in ways that can punish the poor police departments and intelligence services, use big data and algorithms to decide who to investigate

254
00:48:28.150 --> 00:48:30.970
potentially endangering privacy and civil liberties

255
00:48:30.980 --> 00:48:39.629
and authoritarian regimes, linking to the, to the theme of the overall Conference, used the same technologies to repress their citizens and prevent opposition.

256
00:48:44.070 --> 00:48:49.220
In this presentation, I'd like to highlight the role of algorithms in a new domain online fundraising.

257
00:48:49.320 --> 00:48:57.900
This is an interesting case because it's one where private companies use algorithms to provide for needs usually associated with the public sector

258
00:48:58.350 --> 00:49:02.960
like health care and education, thereby blurring the line between public and private.

259
00:49:02.970 --> 00:49:15.819
More generally. This case shows how the spread of algorithms to more areas of social life is linked to transformations in the political economy, like the decline of the welfare state which has implications for poverty and inequality.

260
00:49:18.800 --> 00:49:30.599
So let's start with talking about what is crowdfunding. I'm sure most of you know. But I don't want to assume. So it's you know, what is it? It's it's the practice of asking for money from others, using dedicated commercial internet platforms.

261
00:49:30.940 --> 00:49:40.750
It can be used to raise money for many kinds of projects, and is used in low, middle, and high-income countries alike. Initially it was mostly used for commercial ventures.

262
00:49:40.850 --> 00:49:54.299
like small businesses or new products. But more recently, crowdfunding has become a significant way to fund individual welfare needs. And I'm sure most of you know someone or know of someone who's had a a grant funding page to get help for for something

263
00:49:54.640 --> 00:50:07.100
in our research we call this kind of crowdfunding welfare crowdfunding, which includes fundraising for medical treatments, but also educational costs and for costs associated with accidents in emergencies like fires and storms.

264
00:50:07.150 --> 00:50:13.439
More generally it refers to crowdfunding used to basically ensure a basic level of well-being.

265
00:50:14.670 --> 00:50:24.120
and although welfare crowdfunding happens on many websites in North America, gofundme.com is by far the most dominant service. It operates in 19 countries

266
00:50:24.170 --> 00:50:33.179
in North America and Western Europe, as well as Australia, and has raised over 15 billion U. S. Since its creation in 2,010. This is just gofundme.com alone.

267
00:50:39.030 --> 00:50:48.229
There we go go fund. Me looks like this when you visit the homepage. You see images and descriptions of active fundraisers.

268
00:50:48.290 --> 00:51:00.090
The website also shows you where the fundraisers are taking place, how much money they have raised and what their fundraising goal is. if you find a campaign you're interested in, you can click on it or learn more and donate.

269
00:51:00.570 --> 00:51:07.949
As you can see, the campaigns here have all done really well, and have raised tens of thousands, or even hundreds of thousands of dollars.

270
00:51:08.900 --> 00:51:20.289
Well, there are many issues to discuss about crowdfunding, especially surrounding equity and ethics. What I want to draw attention to in this presentation is the importance of algorithms behind Gofundme and other crowdfunding platforms.

271
00:51:23.850 --> 00:51:47.779
Well, other barriers to successful crowdfunding are also important, especially the digital divide, which we talk about extensively in the in the book. What's less appreciated is how much the platform itself. And this is interesting. Is it the platform, or is it the algorithm? Is it the company who's the actor? And it's a whole? whole? Another theoretical question which I'm now thinking more about, thanks to the last presentation. But it's how much the Platform itself makes decisions about which campaigns succeed in which fail.

272
00:51:48.750 --> 00:51:57.190
Of course, people share fundraisers in their personal networks as well. But the website's interface is an important way for donors to find causes to support.

273
00:51:57.320 --> 00:52:01.429
and, as we can see here, the website directly shapes which campaigns appear

274
00:52:01.460 --> 00:52:08.279
through features like the show, more button. When we click it we see a new selection of campaigns based on algorithmic decision making.

275
00:52:11.710 --> 00:52:18.159
But notice that after we click it, we continue seeing campaigns that are already doing extremely well.

276
00:52:18.380 --> 00:52:33.150
And, as I said, and given the title of the book. This is very concerning, because, as we show and elsewhere in the book, and not really part of this talk, only about 5% of go fund. The users get the help they need on their own terms. So 95% of people do not get the help they need.

277
00:52:33.180 --> 00:52:39.370
So the fact that you're clicking through and seeing more. And then you're seeing successful campaigns is quite interesting.

278
00:52:43.700 --> 00:52:47.720
And even if we search for specific conditions like lung cancer

279
00:52:47.760 --> 00:52:56.219
that we might want to support the results are algorithmically selected so that we continue seeing successful campaigns by demographically fairly similar people.

280
00:52:57.740 --> 00:53:24.620
Now, we should understand that. Go fund. Me does this because it's profitable and go from me as a for-profit corporation. It makes money by taking tips as a percentage of every donation and people are more likely to donate to good looking campaigns that others have supported. So that's why the homepage highlights. These we have a whole section 2, on attractiveness of people, the the quality of the photos, the quality of the text, and how it correlates to success. It's very concerning

281
00:53:27.840 --> 00:53:38.270
using algorithms to highlight campaigns like this is very common on commercial crowdfunding sites as well, like Indiegogo, which you see here, and which is a precursor to go fund me and others.

282
00:53:41.200 --> 00:53:57.220
But we find this practice much more problematic when it's applied to fundraising for filling holes in the social safety net which appears to be the case with welfare. Crowdfunding in this situation being chosen by the algorithm isn't just about whether your business takes off, but it becomes a life or death issue.

283
00:53:57.460 --> 00:54:02.639
What makes us and other scholars concerned is who wins and who gets left behind in the process.

284
00:54:04.770 --> 00:54:15.090
So in this project we don't directly investigate the effect of algorithms. It's hard to do, but particularly for the for Profit company that's not transparent about what's happening with their machine.

285
00:54:15.200 --> 00:54:25.339
but we asked 2 more general questions about how we welfare crowdfunding is linked to gaps in the social safety net, and how equitably crowdfunding can help raise money?

286
00:54:25.490 --> 00:54:30.559
Answering these questions first helps us contextualize the significance of algorithms in this new era.

287
00:54:33.730 --> 00:54:34.750
So

288
00:54:34.770 --> 00:55:01.550
just briefly on the data And this is where Martin was amazing and did lots of heavy lifting. I'm not a computational sociologist, but he is, and we collected over 2 million Gofundme campaigns through the the back door through the site map In addition to this, 2 research assistance manually coded a random sample of 1,000 cases, and also we conducted a ton of interviews. And the the top leader of the research team Carmen is in in the back there, actually,

289
00:55:01.640 --> 00:55:09.260
And so the work we're presenting today primarily involves descriptive analyses of how common crowdfunding is by location and demographic characteristics.

290
00:55:12.080 --> 00:55:33.950
The first set of analyses compares welfare. Crowdfunding usage across different national contexts. So remember, there are first research questions about whether crowdfunding is linked to government. Social supports. These supports differ across countries and regions. For example, the us versus Canada, which we talk about in the book extensively, and so we would expect to see more crowdfunding use in context with weaker welfare, states.

291
00:55:38.110 --> 00:55:44.920
this figure shows bars indicating the per capita crowdfunding campaigns in our data in different countries that go fund me operates in.

292
00:55:44.970 --> 00:55:53.929
and we clearly see that crowdfunding uses by far most common in the Us. Followed by Ireland, Canada, Great Britain, and Australia, and in other Western European countries.

293
00:55:53.990 --> 00:56:02.949
Importantly, this pattern follows our expectation. Given much weaker public supports for welfare costs. And, for example, the Us. Compared to Canada and France.

294
00:56:06.060 --> 00:56:15.540
And it's possible that there are other national differences here other than social policy context. So to deal with this, we can also compare regions within the same country as as we do here.

295
00:56:15.620 --> 00:56:23.189
Here you can see per Capita crowdfunding use in different provinces within Canada, with lighter regions showing more crowdfunding.

296
00:56:23.270 --> 00:56:40.149
we see that the least crowdfunding by far happens in Quebec, and this is consistent with our expectations, because scholars have shown Quebec to have a distinct welfare system that's much more generous and comprehensive than the other English-speaking provinces looking more towards France for inspiration. It's something more like the Scandinavian model.

297
00:56:40.660 --> 00:56:53.939
And together these analyses suggest that social policy contexts indeed shape welfare, crowdfunding usage as we suspected and that algorithmically controlled websites like Gofundme are filling in areas where government supports have weakened

298
00:56:56.620 --> 00:57:09.670
our second set of analyses, compares welfare crowdfunding usage across different demographic groups in the Us and Canada. We're interested in this because research on the digital divide tells us that socioeconomic differences between groups

299
00:57:09.790 --> 00:57:17.510
translate into differences in Internet access and skills, meaning that some are less able to benefit from crowdfunding than others.

300
00:57:20.900 --> 00:57:37.989
And indeed, we see great differences between demographic groups in terms of how likely they are to use welfare crowdfunding, for example, whites compared to Black and Hispanic groups. whites use go fund me in numbers disproportionate to the percentage of the population. I don't have time to unpack all the numbers, but it's pretty dramatic

301
00:57:37.990 --> 00:57:51.370
and you see it with some anomalies. You see, it consistent across the Us. And Canada. Black Canadians, black Americans, indigenous Canadians are just not using go fund me in the numbers that whites are relative to the population

302
00:57:51.510 --> 00:58:09.829
in the Us. And Canadian context. This means that already relatively advantage groups are primarily benefiting from this welfare support in much greater numbers than we would expect, based on their population share in other analyses. We also see that older people are underrepresented. And then there's the digital capacity and the digital divide, and that that matters as well.

303
00:58:12.240 --> 00:58:13.340
So

304
00:58:14.430 --> 00:58:31.969
this is a brief presentation of Of a, of a whole book project. And and There's there's actually a great chapter in there to make a plug about the sort of history and rise of crowdfunding at where it comes from and historicizing it. And whatnot which links to sort of the the business connection. and and the social. But

305
00:58:32.100 --> 00:58:33.020
you know.

306
00:58:33.900 --> 00:58:44.360
we've argued briefly that algorithms play an important role in shaping visibility and success in welfare crowd. I mean a case that blurs the lines between public and private sectors.

307
00:58:44.420 --> 00:59:03.860
And second, we've shown that welfare crowdfunding uses linked to the strength of Welfare Streets, and that welfare crowdfunding in the Us. And Canada is primarily used by privileged groups. So together, this means that as welfare States have declined, private companies have come to play an important role in funding services once mostly funded by the State.

308
00:59:04.030 --> 00:59:12.829
and that algorithms now play an important role in access to and and play a gatekeeping role, really in access to support for health care, education and emergencies.

309
00:59:13.840 --> 00:59:35.040
So we hope to in our next project analyze algorithms more directly if we can. If we can get at them, it's it's tricky. But for now this has concerning implications we know from other research that algorithms often end up punishing people who are already poor and marginalized, and we expect that the same will be the case here, and indeed, finding from our other regression analyses support this.

310
00:59:35.770 --> 00:59:57.259
And we've shown that ethnic minorities and older people are are already less likely to use welfare crowdfunding in the first place, and we believe that algorithms will only make these disparities worse by rewarding more privileged campaigners. What better Internet skills and more attractive campaigns, thereby worsening poverty and inequality. So the the point about algorithms as part of the broader case we make in our forthcoming book.

311
00:59:57.260 --> 01:00:11.309
where we argue that welfare crowdfunding is is really far from the best society has to offer for solving social problems. Most importantly, because it's highly unequal, and most people using it simply failed to raise the money they need.

312
01:00:11.420 --> 01:00:37.029
and then I'll I'll just add a little a few observations. And I I I was pleasantly surprised to find the connections across the talks. I'm a theorist at heart, and and so you know, we board you looms large in in this book for us and and the questions of social trajectory, and who who ends up asking for help in this particularly public way? The data becomes owned by Go fund me. There's all kinds of horrific stuff that goes along with it.

313
01:00:37.030 --> 01:00:50.949
So Borgia is very much present, and the whole impetus for the project was was for me about absence. So I heard of a successful case. A formerly homeless Black Canadian woman who raised money to go to Harvard, which is amazing.

314
01:00:51.090 --> 01:00:55.399
But I found myself cynically, sociologically asking, Who's not asking?

315
01:00:55.630 --> 01:01:25.200
And that was the impetus, who who's not? Who's not crowdfunding? And then there's a twist which for Doris Miller, it might be interesting is that some people have friends make campaigns for them without their knowledge, and so their presence becomes. you know, their digital presence is created, and they don't even know about it. And this can be for very personal things like testicular cancer, or and and some of our respondents told us about this and the really complicated emotions that go through that.

316
01:01:25.200 --> 01:01:53.780
And then you know the multiple purposes from the for the talk before mine, the multiple purposes of of the algorithm. We have profit competing with helping. You know that trying to help people make their goals, but I think it's it's irreconcilable, almost, and I think we we've shown this. And I I expect we'll probably get sued for it. But but who knows? We'll see. But in any event, it's it's it's really nice to to see that sort of, you know. People thinking from different directions are all coming together on the same theme. So with that, I'll I'll say thank you very much for your time.

317
01:02:00.170 --> 01:02:01.670
You about Eric.

318
01:02:02.070 --> 01:02:06.910
I would like to invite Karen and Eric to come into the front

319
01:02:07.450 --> 01:02:09.909
and then I will

320
01:02:10.370 --> 01:02:15.379
go back to the zoom screen so that everybody can see the online

321
01:02:16.200 --> 01:02:18.470
attendees.

322
01:02:18.980 --> 01:02:21.069
I also see some chats

323
01:02:23.080 --> 01:02:24.859
if there is already.

324
01:02:28.640 --> 01:02:32.739
So I think we have everything set up. I mean.

325
01:02:33.460 --> 01:02:40.600
online caneros, too. Right? Yes, I think that should be it. that for the sheriff's.

326
01:02:44.150 --> 01:02:44.850
this one.

327
01:02:47.930 --> 01:02:49.480
maybe for a step.

328
01:02:52.480 --> 01:03:22.139
So who wants to kick off this this session? With a question I can imagine that this most of recent questions. Yes, okay, thank you so much. What a fascinating session my question is for a the colleague who is joining us via zoom a juice to, so I don't know if I'm pronouncing your name properly. Apologies. my name is Barbara. I'm from one Ash University via University of Toronto as well.

329
01:03:22.140 --> 01:03:30.419
and my question to you is, we've been working on. We've published a paper, but we have a project on resistance to AI.

330
01:03:30.420 --> 01:03:55.399
And in our study, what we did was we compared the A-level algorithm in the Uk, which was a failure with the robo that algorithm in Australia, which was also a failure. And so our study was about, how do we resist this kind of welfare based automated service systems in this case and services in your research. Was there any idea of how

331
01:03:55.400 --> 01:04:05.580
is there any resistance to this kind of algorithms in in terms of public sector processes? Thank you.

332
01:04:05.930 --> 01:04:09.510
Thanks to the organizers for a great session as well.

333
01:04:10.800 --> 01:04:34.160
Justus Rahn: Yeah, thank you very much for the question. I hope I understood you right. And I think I mean. What you mentioned was very interesting. It failed to New Case and in Australia. And we didn't mention this because we focused very much on the discourse in the presentation right now, but also in Austria. It I mean it kind of failed. It was stopped by by the role

334
01:04:34.160 --> 01:04:43.449
Justus Rahn: a your restriction and The reason was that it didn't pass data protection law.

335
01:04:43.710 --> 01:05:14.079
Justus Rahn: And there was a lot of resistance, obviously, because it was scaring the the public right? And we, I mean, we had this very niche focus in our research to look on how the observ organization manages to contour the the algorithm discursively, and in that they pretty much succeeded. But there is another form of a power, so to say, and this is

336
01:05:14.080 --> 01:05:22.690
Justus Rahn: the power of law which in that case stop the algorithm, but also that the story is not finished yet, because there is still

337
01:05:22.760 --> 01:05:36.729
Justus Rahn: conflict of law in which they are going through all the instances at the moment. So I'm very interested. If you could tell me your name again I didn't get it the first time, and I would look up your pick.

338
01:05:36.790 --> 01:05:43.790
I'll send it to you. I have a very complicated name.

339
01:05:54.900 --> 01:06:08.879
Brad West University of South Australia. Apologies for being late. I left my party ticket at home, and I had to get back on the tram and get it. So so my questions were, Eric, because that's so sorry. The only paper I heard.

340
01:06:09.590 --> 01:06:17.569
and you know it is a sort of theoretical one, in a way. So you know, you talked a lot about the

341
01:06:18.060 --> 01:06:29.570
go fund, me and and other sorts of, you know, algorithmic funding websites. Yeah. Emerging out of the decline of the welfare state.

342
01:06:30.080 --> 01:06:33.300
I wondered, like, you know, while there is a

343
01:06:33.620 --> 01:06:46.010
correspondence between those 2 things they have the same time, you know. I'm not sure that there's a kind of causality I'll be really interested in if there's empirical data which kind of indicates that that there is and

344
01:06:46.840 --> 01:06:54.719
I've never been. I go fund me page at all. Like, you know, I'm a lud-eyed. I'm sort of quite unfamiliar with it. But I do watch this

345
01:06:55.020 --> 01:07:01.530
 Youtube Australian person about electric vehicles.

346
01:07:02.080 --> 01:07:06.940
and he has a Gofundme page for his wife.

347
01:07:08.040 --> 01:07:14.519
who's suffering from cancer. And but she was told by, you know.

348
01:07:14.630 --> 01:07:17.439
it wasn't a a lack of the welfare state

349
01:07:17.510 --> 01:07:19.150
providing for her condition.

350
01:07:19.370 --> 01:07:22.949
But they rejected the terminal diagnosis

351
01:07:23.970 --> 01:07:33.259
that the Australian health system provided, and the Gofundme is for sponsoring experimental medical procedures in Thailand

352
01:07:33.500 --> 01:07:43.229
which seems to be really interesting, because it's actually, you know, it is the decline of the welfare state, but not in a monetary way, but in in in a faith in and difference to it.

353
01:07:45.730 --> 01:07:58.880
Thanks, Brad, it's It's really interesting. This is a conversation that Martin and I had over months and months and months, and our our first May empirical chapter talks about this and

354
01:07:59.070 --> 01:08:14.050
specifically says it's it's correlated, but we don't see it as causal at all. We're more just sort of charting the rise of crowdfunding the why of it, I mean. And this is where the Borgio comes in. There's sort of this, this.

355
01:08:14.160 --> 01:08:38.940
taking a something that was from the economic field and and taking it without being reflexive about it and putting it into to the field of social welfare and not remembering where it came from. It's almost like the to get theoretical that's thicker than blood to go to vary about statistics, and we don't think about where statistics came from, and it actually was Eugenics movement. And we need to be cautious. So that's that's really where we're at. And for me. The impetus was that

356
01:08:38.940 --> 01:08:51.679
I wrote a book on a nineteenth century charity in North America, in England, and this stuff felt like the same thing again. And I just asked myself, why are people having to beg

357
01:08:51.680 --> 01:08:53.680
on the Internet publicly

358
01:08:53.680 --> 01:08:57.999
in order to get help for things that seem pretty reasonable

359
01:08:58.090 --> 01:09:01.340
and like the kinds of things we should help people with.

360
01:09:01.470 --> 01:09:06.849
and it just felt it felt like the kind of stuff that was rejected in the nineteenth century

361
01:09:07.660 --> 01:09:31.990
for a number of good reasons. Why is it possible to keep track of, you know who's getting help, and whether or not there's fraud there isn't a lock up in the some. to it! It's it's humiliating for people. And it's just it's why social work emerged because of in response to this contract. So so for me. Why, the hackers is coming back again, and more and more and more. And what really bothered me was this sort of rhetoric.

362
01:09:31.990 --> 01:09:40.590
democratization and transparency and and fairness, and anybody can do it, and you can. And it's being

363
01:09:40.740 --> 01:09:42.620
mostly promoted

364
01:09:42.630 --> 01:10:08.770
through the company, which is a for profit company owned by venture capitalist, which makes hundreds of millions of dollars a year on a tip structure that's hidden in dark patterns that you can't even see. It's like, you know some apps when you what you have to off out of tipping or out of things, otherwise it's in there. And so most people don't know it's it's for profit. Most people don't know that there's a tipping system built into the back end for donations, and it feels really bad.

365
01:10:09.150 --> 01:10:10.410
Is it

366
01:10:10.480 --> 01:10:14.380
filling the the gaps in the welfare state definitely.

367
01:10:14.410 --> 01:10:15.639
Is it?

368
01:10:16.030 --> 01:10:24.470
Is it functional, you know? Did it emerge in a sort of functionalist way to? I don't know, you know, that's that sort of big meta

369
01:10:24.960 --> 01:10:28.450
sociological stuff that I can't really, I can't really.

370
01:10:28.460 --> 01:10:29.530
I mean, just

371
01:10:29.730 --> 01:10:50.960
very briefly, I mean, okay, I feel if you assume that you know Canada has more of a welfare state, you know, in in in than the Us. I mean, do you do in your comparative data. Do you see any differences

372
01:10:50.960 --> 01:11:00.730
it up? There's a lot less of it, absolutely. We see it. And you see it. For cancer is a big thing for health care across the board, and there is a lot of stuff for

373
01:11:00.730 --> 01:11:23.119
for marginal, experimental kind of work. But you, you see the differences across. And this is kind of our next project. I just got a gr to look at, to to across nationally, and to look at sort of attitudes about natural zoom attitudes about charity. Look at welfare, state regimes, and almost do like a mills method of difference. Kind of first pass on it. But yeah, that's really interesting. I can sort of.

374
01:11:26.590 --> 01:11:33.519
I have a question. I have a question to Dormicilla. I hope I pronounce well your name.

375
01:11:33.820 --> 01:11:38.059
but I think it relates also to 2 distributed papers

376
01:11:38.430 --> 01:11:42.790
one relates to

377
01:11:42.960 --> 01:11:50.469
retracing algorithm how digital social research methods can trace algorithmic functioning.

378
01:11:50.550 --> 01:11:54.419
And the second one talks about bias as a social issue.

379
01:11:54.650 --> 01:11:57.109
So I was wondering.

380
01:11:57.250 --> 01:12:02.489
no, missila, don't miss it. I'm sorry when you talk about absence.

381
01:12:03.050 --> 01:12:08.879
Do you relate to the exclusionary nature of some algorithms

382
01:12:09.180 --> 01:12:12.469
that the database that they use

383
01:12:12.890 --> 01:12:16.920
is not heterogeneous, and it excludes

384
01:12:17.240 --> 01:12:23.440
 many groups of people who their traces are not there.

385
01:12:23.450 --> 01:12:27.310
and still, when they, when there is a decision making

386
01:12:27.350 --> 01:12:29.090
process made

387
01:12:29.310 --> 01:12:37.779
 there can be a false positive, or a false negative in the person that is doing this decision making, because

388
01:12:37.920 --> 01:12:47.640
there are no traces of those groups. I hope it's you understand my question, or that I so I asked it in a proper way.

389
01:12:51.320 --> 01:12:56.860
Dorismilda Flores-Márquez: Thank you so much for the question. It's Doris Milda, my name.

390
01:12:56.870 --> 01:13:14.720
Dorismilda Flores-Márquez: And well, when I talk about absence in in this case, it's just an exploration, because I'm a trying to find the ways of explore qualitatively a different types of absence in in this kind of

391
01:13:15.210 --> 01:13:17.470
Dorismilda Flores-Márquez: process.

392
01:13:17.520 --> 01:13:22.220
Dorismilda Flores-Márquez: what you say. I think it's one type of absence?

393
01:13:22.280 --> 01:13:30.860
Dorismilda Flores-Márquez: because there are a differences. For example, the most evident. The absence

394
01:13:31.170 --> 01:13:37.689
Dorismilda Flores-Márquez: is a link, it a to the digital divide people that is

395
01:13:38.010 --> 01:13:43.480
Dorismilda Flores-Márquez: excluded because they they don't have access, or

396
01:13:43.520 --> 01:13:44.660
Dorismilda Flores-Márquez: the

397
01:13:44.700 --> 01:13:57.359
Dorismilda Flores-Márquez: capabilities or a different conditions to access, use, express themselves on the Internet. But there are another type of absences that

398
01:13:58.160 --> 01:14:02.339
Dorismilda Flores-Márquez: and has to do with a

399
01:14:02.510 --> 01:14:05.650
Dorismilda Flores-Márquez: this invisibility.

400
01:14:05.680 --> 01:14:13.050
Dorismilda Flores-Márquez: For example, when there are some in hashtags from activists that

401
01:14:13.250 --> 01:14:17.239
Dorismilda Flores-Márquez: a surprisingly disappear or

402
01:14:17.250 --> 01:14:18.860
Dorismilda Flores-Márquez: are a

403
01:14:19.770 --> 01:14:24.820
Dorismilda Flores-Márquez: displaced by a another training topic, or something like that.

404
01:14:25.770 --> 01:14:28.789
Dorismilda Flores-Márquez: There, there are another kind of a absences.

405
01:14:30.180 --> 01:14:36.530
Dorismilda Flores-Márquez: So in, as I said a previously, if I not

406
01:14:36.960 --> 01:14:39.310
Dorismilda Flores-Márquez: answer in

407
01:14:39.540 --> 01:14:43.430
Dorismilda Flores-Márquez: what you are in searching for?

408
01:14:43.540 --> 01:14:46.129
Dorismilda Flores-Márquez: It's because of the night.

409
01:14:52.710 --> 01:15:07.879
Okay, Hi, my, I'm Sam from the Philippines and home under my P. In Hong Kong. My question is for Karina. I'm also actually doing my Phd. On social media influencers. So in the context of the Philippines. But

410
01:15:07.880 --> 01:15:28.889
I'm quite curious. I don't know if I didn't get it in the slide. But or did it mention it? Is there a specific reason for choosing or focusing on you, on Instagram or and is there like a different context with with like climate activists on other platforms in in the Australian context? Thank you.

411
01:15:29.070 --> 01:15:31.769
Yeah, thank you for the question.

412
01:15:31.890 --> 01:16:01.059
Instagram, there's several reasons, because my, my, my case, that is situated in Germany and in Syria at the moment is the most used platform there. across age groups. Of course, if you look at the younger generations like the younger you look at the the more. Tiktok is is popular. but then also, like as it's a Phd, I have like time limits, ethical limits. So like, you know, 18 is a bit more difficult

413
01:16:01.060 --> 01:16:13.440
and then tik tok. It doesn't allow a private messaging functioning. So approaching participants is a bit more difficult. So that's like mainly practical consideration. So I'm going to Instagram to be honest.

414
01:16:13.440 --> 01:16:31.259
But then some of my participants do have platforms or other profiles. For example, on Twitter. Some are also on Tiktok, and we also to talk about that into in the interviews. So I ask them about their activities on these platforms, how they and how they default. How? How? How they relate to Instagram.

415
01:16:38.200 --> 01:17:04.859
Hi! Sorry my voice is a little bit off, but the question is for Eric. I guess. My! I was wondering the role of algorithms beyond the platform itself, because in my imagination, or from from the people I know, I think when people see, go, found me is usually spread through, not directly through the platform, right usually through other social media or the personal network. And so I wonder, do you have data on how this

416
01:17:04.910 --> 01:17:06.380
this

417
01:17:06.470 --> 01:17:24.580
pages are spread, right? Because I think that's kind of interesting part. It's hard to imagine how spread it is that people can raise like millions of successful cases. And what rural does other kinds of algorithms or non? All the reason vectors in those kind of.

418
01:17:24.600 --> 01:17:30.420
And you're inquiring a spreading process. Sure, that's a great question. And the short answer is, no, I don't have any data.

419
01:17:30.640 --> 01:17:59.819
again, I I think it would be really hard to to get at it. And it is. It is this question of, you know, getting at the algorithm itself, I mean, for sure, if the the campaign takes off or blows up or goes viable. Something usually external has happened. The case in the book with Taylor Swift gave $5,000 to a to A, to a person in Tennessee, all of a sudden, $50,000!

420
01:17:59.900 --> 01:18:11.920
It's 24 h and a week when it was on Instagram that was on Facebook. And absolutely, we do some work with Mark Graham's theory on strength of ties and on buttons like breaching capital

421
01:18:11.950 --> 01:18:29.890
people who have better networks do better because they're getting this story out there and then other algorithms are picking up and promoting it. It's a self-fulfilling prophecy, which is the problem because the the ways in which these things get to a front stage, in the first place, are because they seem profitable.

422
01:18:29.890 --> 01:18:45.320
So the algorithm, because the algorithm for go-funding. We believe we don't know what we believe is putting out front those cases which are most likely to get donations, because the only way for funding makes money is by getting donations.

423
01:18:45.520 --> 01:18:56.709
So and that's why, as I showed you, you only see the profitable. But 95% of people don't get the money they need. And I mean, like, really get almost nothing. So

424
01:18:56.760 --> 01:19:01.800
so yeah, I would be fascinating to study that. And I think that's what's so exciting on this panel

425
01:19:01.830 --> 01:19:29.140
and about the the the work that we've done, because it's so generous. There's so much more work to do. I'm hoping that maybe we'll induce go funding to be more transparent about. It's a bias. It's day that I've done room, and we'll see. But I present no, but I need to write the algorithms. It compounds it. You know, a whole lot of limits interacting with one another and either extending or repressing people's stories.

426
01:19:31.610 --> 01:19:33.040
Who's next?

427
01:19:36.170 --> 01:19:47.430
Thank you for the talks. Yeah, I'm Phd. Students in Selji, in Germany, in France. And I wanted to ask a but maybe progative question about

428
01:19:47.630 --> 01:20:08.299
why you're talking about algorithms and not about social phenomenon, or like a community gathering together or filter functions like phenomenon that are already here without algorithm like the rich Getwicher, or in case of the Ams algorithm, we know that it is

429
01:20:08.300 --> 01:20:29.109
wired inside the algorithm that women have will have less chances to get a new job than men. So this discourse and these elements are already here. And then really just a way to maybe put in up front. So I would like you to have your opinion on it. where is actually the algorithm? And what are the

430
01:20:29.180 --> 01:20:30.880
algorithmic impacts?

431
01:20:36.000 --> 01:20:42.160
This is for everyone. So whoever wants to start

432
01:20:42.230 --> 01:20:47.350
also online? please feel free to unmute yourself if you have some. reflections

433
01:20:51.460 --> 01:20:54.980
maybe start because that's

434
01:20:55.410 --> 01:21:06.720
kind of something where I was getting it. Also within my presentation. It's like, well, the algorithm is there, and the algorithm is reproducing these inequalities that are already there in the offline world.

435
01:21:06.730 --> 01:21:15.079
But then this whole idea that I discussed with this algorithmic meta-capital.

436
01:21:15.110 --> 01:21:32.879
And it's the idea of like kind of re-emphasizing this human agency again. And Ily functioning and like finding ways of How do you like it? It's kind of a continuum between like human control and algorithmic control. It's never that it's like fully algorithmic or fully

437
01:21:32.940 --> 01:21:39.850
humanogenic. So it's always somewhere on this line. And I guess.

438
01:21:39.880 --> 01:21:53.940
finding ways and like understanding the algorithms that are. And like, maybe getting companies to be more transparent about the algorithms. But then enable us humans to be more aggressive when interacting with these algorithms.

439
01:21:54.080 --> 01:21:55.639
That will be my applause.

440
01:21:57.540 --> 01:22:02.050
So you're right. And it's sort of like talking about

441
01:22:02.380 --> 01:22:08.859
mass shootings in the Us. And talking about the comments, and not that, you know, but at the same time

442
01:22:09.110 --> 01:22:34.959
I was at a Commemoration memorial session for Martin Archer. And we were talking about AI, and I mean algorithms are very shortly going to to be more than just tools. And and that's it. That's a scary thing to think about as well at the end of the day. This is this is the same. These are the same social forces, the same capitalist excess, the same the same big social problems that we've been experiencing. But it's package

443
01:22:35.000 --> 01:22:46.610
in a very clever, sneaky way to make us think it's something democratic and fresh and new. So there's a there's a slide of hand going on for me and I.

444
01:22:47.070 --> 01:22:50.689
And so to focus on the tool, I mean.

445
01:22:50.950 --> 01:23:17.419
And that's I'm really thinking about this a lot like, who's the agent here, is it? Is it the programmers? Is it the you know, who's complicit in the story? But for sure, this is just the latest expression of of things that we could see for a long time like I said, what I what I saw in nineteenth century chair doing historical work I'm seeing again now. It's the same stuff. There's no difference at all.

446
01:23:18.180 --> 01:23:40.949
I I think there is actually machine learning, I mean, from what I understand. But I you know we but this is the problem we don't know. But what concerns me is that there's a massive black box that surrounding tens of billions of dollars, of what should, in my view, be social welfare, assistance. And so, you know, when you're talking about tuition and cancer and and these kinds of things, we we want to know.

447
01:23:41.080 --> 01:23:46.319
and really I mean the punch on me. I don't think we should be doing this at all.

448
01:23:46.850 --> 01:23:52.530
We should be happy for you through food speed. But that's the I.

449
01:23:52.740 --> 01:23:56.419
But it's a great question. You're absolutely right to ask that.

450
01:23:58.100 --> 01:24:12.019
Yeah, I was some muting already. thank you very much. I mean, you're right. We didn't explicitly focus on the algorithm

451
01:24:12.020 --> 01:24:35.930
Justus Rahn: as a research object. There are a couple of reasons. I'm trying to cut it short. So first, at least in the German discourse. It's very stark focus on technology algorithm. The techno sociological focus is very strong. And our no, our our observation was that the organization as a research

452
01:24:36.260 --> 01:24:58.719
Justus Rahn: object in the discourse about digitalization and algorithms is often neglected. So we wanted to feel that this very little gap in which we don't. Really, I mean, you're absolutely right, right? The algorithm as is perpetuating this, the the discrimination towards women towards people of color that is, that they are

453
01:24:58.990 --> 01:25:03.250
Justus Rahn: lower ranks in the algorithmic decisions.

454
01:25:03.300 --> 01:25:20.569
Justus Rahn: And the interesting thing is that the organization succeeds in framing this algorithmic tool as tool right in in which they say, Yeah, I mean, you get a point there. But the algorithm does nothing that the

455
01:25:20.780 --> 01:25:27.920
Justus Rahn: that the employee of the Ims. The Ims Counselor will not do him or herself as well.

456
01:25:28.420 --> 01:25:32.650
Justus Rahn: and I mean, I think, that that this was

457
01:25:33.050 --> 01:25:46.280
Justus Rahn: a fruitful observation to. To also put into that the organization has a massive power to frame the algorithms functions and

458
01:25:46.640 --> 01:25:55.129
Justus Rahn:  and sort of effects in the way they actually did, and that I was successful on the public level.

459
01:25:55.390 --> 01:26:14.159
Justus Rahn: Alright. So I I get your point. I think. In that case. there's a lot of research on on data and studies. And we would like to sh, or we were like keen to shift the focus to the role that the organization plays in the discuss.

460
01:26:20.680 --> 01:26:22.950
Justus Rahn: Sorry. Does that answer your question more or less?

461
01:26:23.310 --> 01:26:31.019
He's saying yes, he's saying yes.

462
01:26:31.580 --> 01:26:36.960
and if Doris Miller, you want to respond otherwise, I I have a question. No, okay.

463
01:26:37.160 --> 01:26:41.030
so this is a Iila Martin, speaking. Thanks a lot for

464
01:26:41.190 --> 01:27:02.810
for your presentations. Really enjoyed every one of them very fascinating. The diversity of topics that were covered from Instagram influences all the way to the large, you know, societal structures and inequalities, and the more philosophical absence presence issues. So I really love that because I think it's necessary to have that broad variety

465
01:27:02.950 --> 01:27:27.860
and also following up the keynote a speech we had yesterday. I was in Jeremy. you you brought this up to have digitalization as that which needs explanation. I think we social. I just have a key responsibility to open that up for sociological research and not leave it to technologists, to and technocrats to to run that. Having said that, My question, it's a follow up actually is.

466
01:27:27.920 --> 01:27:30.129
is, in a sense.

467
01:27:31.560 --> 01:27:33.619
Did you have to

468
01:27:33.890 --> 01:27:44.080
an update, modify, change your sociological instrumentarium to make sense in particular because you're doing you, you're doing with algorithms.

469
01:27:44.500 --> 01:27:51.640
Or did all those concepts that we have been developing for industrialization and modernity

470
01:27:52.340 --> 01:27:57.830
work for you just as well, so in other ways do we require to

471
01:27:57.950 --> 01:28:03.809
upgrade our sociological thinking because there is something specific about algorithms.

472
01:28:03.820 --> 01:28:06.889
It's not just more of the same, which I think, is also very, very true.

473
01:28:07.130 --> 01:28:18.629
But is there something specific that because you're dealing with algorithms that allowed you to understand something about your particular phenomenon in a very different way that made a a change that would also require

474
01:28:19.140 --> 01:28:25.710
changing our way of sociological thinking. For instance, agency, for instance, I think, is not a helpful way to look at this here.

475
01:28:29.510 --> 01:28:43.290
 yeah.

476
01:28:43.980 --> 01:28:52.739
Yeah. I also attended the keynote yesterday, and I definitely am interested to buy the book and then more about it.

477
01:28:53.860 --> 01:29:02.539
 but yeah, I do think, especially in the cycle of social media platform environment.

478
01:29:02.670 --> 01:29:04.260
Yes, it is like

479
01:29:04.610 --> 01:29:06.390
privatized.

480
01:29:06.960 --> 01:29:13.769
I square where we are all in acting, and it's like, I guess most of us.

481
01:29:13.910 --> 01:29:15.260
And then

482
01:29:15.630 --> 01:29:18.939
that is new. That is new.

483
01:29:19.080 --> 01:29:21.109
unbounded kind of

484
01:29:21.160 --> 01:29:32.319
spaces where? Yeah, where we can try to to use all of that like more traditional like theories and ideas.

485
01:29:32.420 --> 01:29:33.850
But

486
01:29:34.060 --> 01:29:46.129
I guess. Yeah, we do need to maybe rethink them, and also like the one that I've been using the little he was trying that, in a sense. But yeah, I'm curious also to

487
01:29:52.540 --> 01:29:59.169
you know, the theorizer. Just let go. And that's funny. It's it.

488
01:30:00.620 --> 01:30:21.620
I'm working on a paper which I'm calling, re theorizing the crowd, which was a big question, is this the crowd the same as it was, you know, Park, or you know the ball, or going back, and yes or no. But I don't know the answer honestly, I'm still sort of trying to figure it out. in in this context. And and others.

489
01:30:21.960 --> 01:30:47.750
I think the short answer would be not really. I think we did fairly well with with existing sociological theory, primarily, because we were looking at at at outcomes and motivations, not at the at the machine inside, as much to the extent that we were looking at the machine inside, and sort of Microsoft. And what have you? But I think as it as

490
01:30:47.750 --> 01:31:06.250
and I and I, you know, a filter algorithm is certainly an algorithm in some senses, and we're co-funding. I mean, I know this from from reading. But you know, in terms of anticipating what's going to be profitable and making decisions on the part of the company. It's not just because it's impossible to navigate your way. It's not. It's a frustrating constant.

491
01:31:06.400 --> 01:31:13.510
you know, putting in front of you, driven by something back there. What is that? I don't know. I don't know how to understand it.

492
01:31:13.510 --> 01:31:36.730
I don't know. I mean, there's the garbage in garbage out weapons of math destruction, that kind of stuff. But I don't think as theorists. We really have any idea what that is. And it seems like it's going to matter more and more. So that's sort of where I'm most interested in going theoretically going forward myself. So I think it's a it's a great question. It is that question truthfully for for us thinking about this and stuff.

493
01:31:37.170 --> 01:31:38.729
A lot of the others have to say

494
01:31:40.780 --> 01:31:44.790
online, would you like to reflect on this.

495
01:31:45.350 --> 01:31:46.190
or is

496
01:31:47.920 --> 01:31:51.100
Yes, no, please go ahead.

497
01:31:52.220 --> 01:32:00.070
Dorismilda Flores-Márquez: Yes, but it's it's about the the previous question about this need to to to the think

498
01:32:00.140 --> 01:32:07.860
Dorismilda Flores-Márquez: and the sociological perspective, or something like that. I I think this kind of objects.

499
01:32:07.930 --> 01:32:18.449
Dorismilda Flores-Márquez: can it be read from? just one discipline in this case, just from sociology? This, a objects need a interdisciplinary perspectives?

500
01:32:18.480 --> 01:32:29.089
Dorismilda Flores-Márquez: when we talk about a algorithms. And we, we talk about this.

501
01:32:29.460 --> 01:32:36.170
Dorismilda Flores-Márquez: he works this day studies with data, may mining or

502
01:32:36.200 --> 01:32:49.590
Dorismilda Flores-Márquez: a big data or something like that. There is a data science. of course, and in in some qualitative perspectives are

503
01:32:49.880 --> 01:33:00.550
Dorismilda Flores-Márquez: I know this is a go with other disciplines, such as anthropology, a communication in my case and the

504
01:33:01.170 --> 01:33:11.030
Dorismilda Flores-Márquez: political science. So I think this is a, of course, this, this is a a need. We need to think this

505
01:33:11.220 --> 01:33:16.490
Dorismilda Flores-Márquez: perspectives, but in all the disciplines. And then

506
01:33:16.900 --> 01:33:23.809
Dorismilda Flores-Márquez: for in order to to approach the complexity of the objects.

507
01:33:23.830 --> 01:33:26.819
Dorismilda Flores-Márquez: the interdisciplinarity is basic

508
01:33:30.670 --> 01:33:33.270
justice. Any reflections.

509
01:33:37.000 --> 01:33:50.180
Justus Rahn: I have switched off my camera because I am having a little bit of connection issues on my side here. So but I am of, obviously, I mean algorithms.

510
01:33:51.170 --> 01:33:52.430
Justus Rahn: our

511
01:33:52.549 --> 01:34:16.089
Justus Rahn: research objects that have not been explored for for decades and centuries. So there is, of course, and I see this every day in my research project. Apart from Ams, I'm I'm looking into case management systems into a medical test software

512
01:34:16.090 --> 01:34:33.410
Justus Rahn: that that is around algorithmically. And it's really really hard to understand it with the current theoretical, observational, that bratches that we have right. So I am very happy about the Leonardi as a social materiality approach sometimes

513
01:34:33.620 --> 01:34:46.010
Justus Rahn: helps me understanding the implication of various systems, organizational practices human actors that play a vital role.

514
01:34:46.190 --> 01:35:00.160
Justus Rahn: But it's a hard work. I mean. It's kind of nothing that that's has been explored for for quite a bit. It's it's sometimes challenging for me also, sometimes frustrating. But in the end

515
01:35:00.469 --> 01:35:02.499
Justus Rahn: something that is

516
01:35:02.520 --> 01:35:03.950
Justus Rahn: like very

517
01:35:04.030 --> 01:35:14.220
Justus Rahn: fruitful, and it also gains some some liberty. Right? Because if you're if you exploring such new and some kind of fluids

518
01:35:14.260 --> 01:35:21.640
Justus Rahn: phenomena. You are more free to explore them. combining

519
01:35:22.129 --> 01:35:24.889
Justus Rahn:  approaches

520
01:35:25.120 --> 01:35:31.839
Justus Rahn: so might not seem intuitive at the first time. That would be

521
01:35:32.010 --> 01:35:33.939
Justus Rahn: kind of. The answer to that question

522
01:35:39.889 --> 01:35:40.590
was a

523
01:35:40.870 --> 01:36:06.410
Thank you. Can I? Just a comment? And then a quick question. It's just funny, this conversation we're having about new concepts. And because we're doing a systemeric review on resistance to AI. And I'm sorry, Doris Milda, because this is I'm going to say something that is not very kosher. But our colleagues in Median communication every time they talk about algorithmic something they have to come up with a new concept. And then they conflate concepts with theories. And so

524
01:36:06.410 --> 01:36:25.309
I'm actually trying to see if we can reduce the new concepts in the new theories because a lot of what we use in social science and in sociology is already actually quite relevant. That was just my quick comment. Feel free to disagree with me. my quick question. Sorry, Karine. I wasn't here when you started your presentation. So probably I missed this.

525
01:36:25.309 --> 01:36:38.500
Did you address any folk theories how people think that algorithms work. And this is also for Eric, do actually, people understand how the algorithms work? Or is this all folk theories, how they think the algorithms work. Thank you.

526
01:36:41.059 --> 01:37:08.209
So in, in case of my participants. Not that no idea. So they were just assuming assuming what they think they heard from somewhere. One of them she was when she referred to using hashtag. She was like, Yeah, that's how I learned it. Like, that's how I learned to when I started using this platform. But I I don't even know if it's like relevant anymore. So that was her reflection on it. But no one, no one had any idea. Yeah.

527
01:37:08.410 --> 01:37:13.329
they were quite certain about the commuting time. So so that that seems to be a good bet.

528
01:37:13.950 --> 01:37:22.170
So no, I mean we. It wasn't something that we asked about specifically, if you understand the AI. But we asked a lot of questions about the platform, and

529
01:37:22.810 --> 01:37:27.599
most people have no idea about any of it, and they just

530
01:37:27.870 --> 01:37:30.020
think it's some sort of benevolent

531
01:37:30.790 --> 01:37:54.579
place where they can put themselves out there and then their friends and nice, you know strangers can find them and and give them money, and they don't know that it's for proper they don't I? You know I would. I would save a lot of the responses, even though what an algorithm is and they, you know, they don't know if they don't know what dark patterns are. They don't understand. I mean, it's it takes us to a secure user. And this also, we're forgetting that a lot of people

532
01:37:54.680 --> 01:38:09.630
can't even access this. I'm talking about absence, I mean, that's the thing for me, is, there are a number of people in the Us. In Canada, and here, I'm sure, as well, who don't have access to a phone, and you don't have access to Wi-fi, so it doesn't even matter, so they can't even access this

533
01:38:10.310 --> 01:38:22.280
perverted but sometimes helpful system to to get any support. But yeah, I think that's a really great question. That would be interesting, too, you know. Do you understand how this works is that you're using?

534
01:38:22.420 --> 01:38:23.530
Probably not.

535
01:38:23.680 --> 01:38:30.519
I know which pictures will be more if you are saying that some features, actually.

536
01:38:30.570 --> 01:38:53.830
you are saying that some photos actually raise more visibility on the platform than others. Do people actually know? What would they call folk theories they have a sense of.

537
01:38:53.850 --> 01:39:16.439
you know, or or if you don't have good grammar, you're not going to get help. So sort yourself out. I mean, that's obvious for anyone who knows anything about marketing. But most people don't think about that, and it's it's painful to see, because this is about. You know, babies dying in in intensive care, units and stuff. And it, you know. It's these sort of very painful pictures that are.

538
01:39:16.440 --> 01:39:25.380
You know. They're they're poorly done. They're not thought about it from a marketing, and that's what's so disgusting is that you're really saying you need to look at the market. You're selling the problems.

539
01:39:25.380 --> 01:39:31.020
which is what I find so distasteful. But no, most people don't really think about it. In fact.

540
01:39:31.090 --> 01:39:44.380
just the other day of good friends of mine put it, go fund me up to help bail out their business. It's full of typographical errors. It's problematic it. And look at you saying, Oh.

541
01:39:44.730 --> 01:39:58.249
did you not think about it at all? And probably not. is the answer. No one, no one's helping them, I mean, go finally gets enough money off the people who know what they're doing that they don't need. Now, what what interest would it be for a firm to to do that?

542
01:39:58.350 --> 01:40:01.870
I you know. So that's, I think, the the real problem with all of this

543
01:40:03.190 --> 01:40:28.680
sorry, just as I see your hands. But I just wanted to follow up on the question to you, Egg, because listening to you about this, it's also, I wonder is it also because it might feel a bit taboo is to have a have a marketing tactic for a gofundme profile like this? Because if we talk about tinder, for example, we now have a whole list of things that you need to have on a picture you need to. As a man. You need to hold a fish, you know these kind of things, so it's

544
01:40:28.900 --> 01:40:35.790
why not to go from me, then? Well, I think you're you're right. It is somewhat

545
01:40:36.050 --> 01:41:01.730
taboo, but you know there are people who know, and they're usually younger, or they have children who are sophisticated, and they know, like my daughters always tell me, don't take a picture of the square ticket at an angle and do this and don't move back. Use it, and there's a there's a specific additional sophistication to a certain subset of the population that I think infuses. I mean, like I. You know, I I don't think it's even thought of this marketing.

546
01:41:01.730 --> 01:41:10.189
It's just a way, I think younger people present themselves in digital space. You know. My daughter shares her location with, I think.

547
01:41:10.190 --> 01:41:13.369
50 or 60 people from, and they love it.

548
01:41:13.430 --> 01:41:35.890
and they don't think about it at all. And it's the same thing with, you know, making yourself look good taking a lot of different pictures. The copy. I I mean, you know, but it is a problem, and it is somewhat taboo for some people, and that speaks to me and having some sort of not for profit assistance to. If we're going to use this form, it's technology. And we don't want to blame the technology and keep using algorithms and and the digital platform

549
01:41:36.110 --> 01:41:40.219
people need help. And there needs to be some sort of rationalization of the entry

550
01:41:40.230 --> 01:41:48.939
and helping people who can get their other ones. That's a great question. Just as do you want to? still, reflect.

551
01:41:49.520 --> 01:42:01.459
Justus Rahn: Yeah, just to make sure I can. I'm sorry for the sound issues. Was that a question regarding everyone? Or what's that specifically for? That's all. I and I go for me.

552
01:42:01.960 --> 01:42:10.610
No questions are for everyone. So so yeah, I found it very interesting that you asked what people know.

553
01:42:10.690 --> 01:42:12.950
Justus Rahn: how the algorithm works

554
01:42:13.400 --> 01:42:26.030
Justus Rahn: and my answer would be. They don't necessarily know that from a technical standpoint, I mean, this is something that's even some programmers

555
01:42:26.290 --> 01:42:55.500
Justus Rahn: erez agmoni don't understand any more. I mean, rob kitchen reflects on the opacity of big data algorithms and so on. And in our specific case it was very interesting that the organization obviously had an idea how this algorithm works discursively, but from a technical standpoint I think they don't really. And the other discourse carriers, for example, this famous computer professor, who was a very present during the 2 years in the public media. One

556
01:42:55.610 --> 01:43:11.270
Justus Rahn: she had no idea, and she fell in various traps, implying that there was some AI technology which it isn't right. It's it's a pretty simple algorithm compared to many, many other machine learning things.

557
01:43:11.680 --> 01:43:20.969
Justus Rahn: And this was something that had massive winnings in legitimacy for the deploying organization.

558
01:43:21.020 --> 01:43:32.359
Justus Rahn: Obviously they were successful and evoking confusion. And even, I mean, even computer professors failed at

559
01:43:32.630 --> 01:43:47.789
Justus Rahn: clearly articulating what they thought were the issue. So I would say, No, and I think that's an issue. But it's not an issue that can be resolved very easily, because those things are extremely complex

560
01:43:48.040 --> 01:43:56.629
Justus Rahn: and not even in in the developer teams. everybody knows the function. So I I think

561
01:43:57.270 --> 01:43:58.530
Justus Rahn: knowing

562
01:43:58.820 --> 01:44:02.239
Justus Rahn: in terms of folk theory is not the worst.

563
01:44:06.640 --> 01:44:08.059
Other questions.

564
01:44:09.590 --> 01:44:25.760
Well, I have a final question for doors. Me that because I I really like this tension slash, link that you point out between presence and absence. I think it's also very triggering way to think about our own presence and absence on the social media, for example.

565
01:44:25.910 --> 01:44:36.930
Yet at the same time. It's also you point out an Hegelian problem. And if we talk about Hegel, we do not talk about algorithms.

566
01:44:36.930 --> 01:44:59.620
Because if you you know, if you would follow this, then even Hegel would say, to be absent, you need to be present, because to be acknowledged as absent. It means that someone needs to realize that you are absent, meaning, if you are at a party, you need to have your friends there that make people realize that you are not there. So there is this second kind of second order presence that you are dealing with

567
01:44:59.720 --> 01:45:24.409
and that that in your examples that was for me kind of similar. So because your wife is on social media or your husband is on on social media. In the end people realize that you are not, and you are traceable through them. so for me, I wonder how much is then different when we talk about algorithms? And digital? Is it really this

568
01:45:24.410 --> 01:45:31.260
second-order absence. Is there maybe another type of absence that we are not even talking about at this moment?

569
01:45:31.640 --> 01:45:33.770
Do you have any reflections on that?

570
01:45:36.300 --> 01:45:39.040
Dorismilda Flores-Márquez: Yes, or maybe a

571
01:45:39.510 --> 01:45:53.689
Dorismilda Flores-Márquez: different kind of a slides of the of that is express in that in terms of algorithms, but a as well in terms of a decisions.

572
01:45:54.550 --> 01:46:04.440
Dorismilda Flores-Márquez: for example, people that this site a. Not to be in social media? A. But that is it there.

573
01:46:04.460 --> 01:46:08.830
Dorismilda Flores-Márquez: But, for example, people that the late

574
01:46:08.990 --> 01:46:10.310
Dorismilda Flores-Márquez: their content

575
01:46:11.220 --> 01:46:21.190
Dorismilda Flores-Márquez: that the side note to be not to be there. So  I think it's a it deserves.

576
01:46:21.420 --> 01:46:27.140
Dorismilda Flores-Márquez: Further conversation about the implications of a talking about the

577
01:46:27.340 --> 01:46:34.450
Dorismilda Flores-Márquez: algorithmization platform is session theatification, Digital digitalization?

578
01:46:34.500 --> 01:46:40.410
Dorismilda Flores-Márquez: Because, there are some implications that a

579
01:46:40.870 --> 01:46:46.210
Dorismilda Flores-Márquez: a transform our experience in social life

580
01:46:46.380 --> 01:46:54.740
Dorismilda Flores-Márquez: in in the media, but also that transform a the ways we as researchers.

581
01:46:54.800 --> 01:46:58.340
Dorismilda Flores-Márquez: try to follow these changes.

582
01:46:58.980 --> 01:46:59.710
Yeah.

583
01:47:00.190 --> 01:47:07.360
Dorismilda Flores-Márquez: I hope to be. I more or less clear, absolutely. Thank you.

584
01:47:07.590 --> 01:47:12.109
Anyone wanted to ask a final question, share a reflection.

585
01:47:13.350 --> 01:47:18.209
anything? Well, I want to thank all participants

586
01:47:20.140 --> 01:47:27.619
again for the organization of this session. Fantastic presentation and very nice reflections.

587
01:47:27.770 --> 01:47:34.979
That's it, I guess, for today everybody go to the party, have a drink, see you tomorrow.