WEBVTT 1 00:00:00.001 --> 00:00:06.329 hello, everybody. I welcome everybody online. And here in Australia. 2 00:00:06.530 --> 00:00:14.729 to this session about digital discourse in democracy within this. 3 00:00:15.420 --> 00:00:25.120 with this larger session of digital futures. it's a third sub session of this session. And 4 00:00:25.270 --> 00:00:29.490 before starting, I'll start with the acknowledgment of the country 5 00:00:29.970 --> 00:00:35.840 on behalf of everyone gathered. I pay respect of the spirit of country here and Irene. 6 00:00:35.950 --> 00:00:41.179 There we are meeting. I acknowledged the past and present cooperation bone Jerry 7 00:00:41.970 --> 00:00:52.640 as the sovereign people here the cultural knowledge holders who maintain and promote the 9 Arab, but aboriginal law, a gustans and custard custard chip 8 00:00:52.650 --> 00:00:56.879 responsibilities for country and the people of that country. 9 00:00:58.270 --> 00:01:02.240 and continues to sustain generation of the generation into the future of 10 00:01:02.300 --> 00:01:10.960 for all of us. I recognize the important contribution that first nations thinkers have major sociology, including challenges 11 00:01:11.270 --> 00:01:13.720 to our dominant ways of thinking. 12 00:01:13.770 --> 00:01:21.679 We thank all first nations delegates in the room today for their contributions to knowledge and in the discipline. 13 00:01:23.320 --> 00:01:25.979 This is a 14 00:01:26.720 --> 00:01:31.760 unfortunately, some of the people and 15 00:01:38.260 --> 00:01:56.120 The first speaker, Gabriela, who should have recorded her presentation. isn't here, and we don't do not find the recording either, so I ask Marina to start with her presentation. Marina 16 00:01:56.380 --> 00:01:57.830 is 17 00:02:01.870 --> 00:02:08.840 a postdoc research at the universe, Munivzi Daj Estadal, the camp and campus, Brazil. 18 00:02:09.430 --> 00:02:11.310 And she will 19 00:02:11.480 --> 00:02:18.210 give us a presentation about Brazil democracy and deep fakes overseeing the 20 20 00:02:18.740 --> 00:02:22.830 22 elections. Please marry us. 21 00:02:23.020 --> 00:02:24.350 you have the voice. 22 00:02:25.750 --> 00:02:37.339 Marina Fontolan: Okay. Thank you. So both me and equally Michael Walters will be presenting. I'll I'll just share my screen here, and we can 23 00:02:38.250 --> 00:02:39.130 Marina Fontolan: the 24 00:02:40.430 --> 00:02:59.240 Marina Fontolan: just for a quick A quick announcement. This This presentation is part of a larger project called Ender which all about covid the short along third and long term effects of the COVID-19 pandemic, including 25 00:02:59.260 --> 00:03:00.630 Marina Fontolan: the 26 00:03:00.730 --> 00:03:09.219 Marina Fontolan: the elections here. Here, here, here, Brazil, And for more information you can find on our website. 27 00:03:09.460 --> 00:03:13.749 Marina Fontolan: So can you all see my slides here. 28 00:03:17.270 --> 00:03:21.080 Marina Fontolan: Yes, yes, yes, okay, perfect. 29 00:03:21.220 --> 00:03:23.970 Marina Fontolan: So we can. We can begin. 30 00:03:24.640 --> 00:03:28.950 Nicole De March: Yes, so Hello! 31 00:03:29.020 --> 00:03:41.589 Nicole De March: So mine presented herself like he's my colleague in the department of Science Technology policy at the University of. 32 00:03:41.920 --> 00:03:47.429 Nicole De March: So with the our colleagues. 33 00:03:47.710 --> 00:04:00.250 Nicole De March: this is silver in my con, fully by Ros with an analysis. So our research is called Brazil democracy and deep fix. Overseeing 2,022 elections. 34 00:04:00.290 --> 00:04:11.020 Nicole De March: We would like to thank the University of Copenhagen for hosting our research and also fake forapes and copies for the financial support of our research. 35 00:04:15.780 --> 00:04:24.090 Nicole De March: so our main goal in this work also analyze the the fakes and tips. 36 00:04:24.260 --> 00:04:27.559 Nicole De March: use it to zoom for me during to 2,000. 37 00:04:27.620 --> 00:04:43.199 Nicole De March: So it's 2 Alex elections in Brazil. So this structure of our work we start with the definitions of deep fakes to fix. And this information. 38 00:04:43.400 --> 00:04:49.359 Nicole De March: So then we go to the mythology, our of our analysis. 39 00:04:49.610 --> 00:04:56.449 Nicole De March: and we present the quantitative analysis, the qualitative analysis, and then their results. 40 00:05:01.020 --> 00:05:04.789 Nicole De March: So, according to our though in direction. 41 00:05:04.820 --> 00:05:21.140 Nicole De March: this information is information that is false, and the person who is disseminating it knows it's false. It's the liberated intention online in what it wants to. People be actively disinformed by malicious actors. 42 00:05:25.140 --> 00:05:36.479 Nicole De March: So another definition, for this information is giving by my shadow in the 43 00:05:37.210 --> 00:05:43.840 Nicole De March: but they are that these lies. I use it strategically in order to get the advantages of end sorts. 44 00:05:44.160 --> 00:05:46.300 Nicole De March: So first. 45 00:05:46.470 --> 00:05:59.819 Nicole De March: I'm going to define the deep 6 and 56, they're both out. Visual manipulations. Uses to Z for zoom form. So the difference is that deep fake 46 00:05:59.920 --> 00:06:04.429 Nicole De March: use the artificial intelligence and machine learning 47 00:06:04.630 --> 00:06:18.660 Nicole De March: and the cheap fix uses simple innovating tools using software or not. So we'll try and eventually create you mislead in fireplace false evidence of something that never occurred. 48 00:06:18.730 --> 00:06:21.779 Nicole De March: So they're both used to this in form. 49 00:06:26.340 --> 00:06:29.610 Nicole De March: So 50 00:06:31.080 --> 00:06:37.839 Nicole De March: they this all the visual manipulation it includes a lot of techniques. 51 00:06:37.870 --> 00:06:43.139 Nicole De March: So it's both the cutting edges. our feature. And there's a related techniques. 52 00:06:43.240 --> 00:06:55.969 Nicole De March: As I said it deep. Fix as well in cheap fix that using conventional techniques like we start to photos. So Photoshop is still images. But then we can go to video 53 00:06:56.050 --> 00:07:03.549 Nicole De March: like speeding is loading visual coding, rest, Asia or reconstruct contest tailizing for fish. 54 00:07:07.500 --> 00:07:17.929 Marina Fontolan: Okay? So as for the methodology of this research, we decided to take a look into the fact-check to effect. Check a agencies 55 00:07:18.010 --> 00:07:34.620 Marina Fontolan: and we use their search engines with the keywords at least point which means in a elections in English and 2022 we checked on these 3 websites.org, which is more specialized, in 56 00:07:34.710 --> 00:07:38.690 Marina Fontolan: fact, checking information that is circulating on Whatsapp. 57 00:07:38.840 --> 00:07:50.420 Marina Fontolan: And then I'll spot those. And, as I say, Lupa, which are more focused on these more public social media, such as Facebook, Instagram, Twitter, etc. 58 00:07:50.440 --> 00:07:52.680 Marina Fontolan: So 59 00:07:53.910 --> 00:08:12.400 Marina Fontolan: so One thing one thing that we decided to do was to stop at those 3 fact tech agencies, because we saw that like, the information was repeating itself. So it would it bring any new data to our to our research. 60 00:08:12.420 --> 00:08:26.400 Marina Fontolan: and we are also going to use some of the outspoken reports they do. they do publish Annual reports in which they state which kinds of this information we're seculating on that year 61 00:08:26.570 --> 00:08:35.550 Marina Fontolan: and and like the types, and also like the content of what they have changed. 62 00:08:35.730 --> 00:08:50.069 Marina Fontolan: So let's start with the out spot. Those reports So this is the 2022 reports they have checked around the 700 this information pieces in social media. 63 00:08:50.220 --> 00:08:54.220 Marina Fontolan: and here you can see 64 00:08:55.100 --> 00:09:07.029 Marina Fontolan: like a breakdown of these around 700 the information pieces. So if you take a look on your left, Lula is our current President. 65 00:09:07.240 --> 00:09:18.820 Marina Fontolan: And he was a runner up for the elections in 2,022, and also herable sonato, our Firmware President, and also a runner up. So they were the major 66 00:09:18.830 --> 00:09:28.370 Marina Fontolan: things for this information. but if you can take a look at the grayer areas here we have fraudulent to out, which means election fraud 67 00:09:28.400 --> 00:09:41.030 Marina Fontolan: also like manifestations. And our our electronic voting system here in Brazil. So as you can see, there was a focus 68 00:09:41.130 --> 00:09:57.809 Marina Fontolan: on on this information about the election. So when we take a look at Nola, they have checked a hundred 56 citations there. From this the best majority was to a tech luma 69 00:09:57.830 --> 00:10:09.589 Marina Fontolan: saying that she was in like a good candidate and etc., and when we go to Herable, so now we had a hundred 51 citation. So, as you can see, it's around the same amount 70 00:10:09.720 --> 00:10:25.950 Marina Fontolan: of of this information that I was far to head stretch but ninety- of those were defending. So we can see by this data that they were. there was a very clear 71 00:10:26.050 --> 00:10:31.520 Marina Fontolan: which is information this information. 72 00:10:32.650 --> 00:10:47.940 Marina Fontolan: like it. Just information effort from shire, both of mark from Jerry Boson and the far right side. Then, of course, from the left, the center left to right listen to left side. That was 73 00:10:48.070 --> 00:10:51.999 Marina Fontolan: that was well, let me. Of course there was some 74 00:10:52.210 --> 00:11:02.830 Marina Fontolan: some this information from the left, but when we get to the far right, it's in a much in a in a much larger amount. So 75 00:11:02.990 --> 00:11:15.700 Marina Fontolan: when we get to the elect our electoral system. We can see we can see that the same things happens here, at least for what what I'll fax just had checked. on that year. 76 00:11:15.760 --> 00:11:30.739 Marina Fontolan: We can see that like in the beginning of the year. there they started some some this information about our our elections and our authorities in elections. So the 77 00:11:31.380 --> 00:11:41.600 Marina Fontolan: so our justice system had do have a, that is all about the elections. And they were starting to attacking it. And then when we get to October 78 00:11:41.750 --> 00:12:01.299 Marina Fontolan: we have, like a peak of the amount of this information circulating about our about our elect to about our election system, which is exactly when the elections happen. So both first and second terms here happens in the month of October 79 00:12:01.380 --> 00:12:16.409 Marina Fontolan: of 2022, but they started, of course, I mean building this information and distrust from the system. at the beginning of the year, and then when it got to the elections itself, and even after it 80 00:12:16.490 --> 00:12:25.830 Marina Fontolan: they continue, they continue attacking attacking. And these informing people about our systems. So 81 00:12:26.130 --> 00:12:43.989 Marina Fontolan: there is also here, a clear message of what they wanted to do. So you distrust of people about what was happening. and of course from there we can talk about how they exactly did that. So 82 00:12:44.380 --> 00:13:08.609 Marina Fontolan: so the vast majority of those weren't like sheep or deep fake situations. So we brought here some examples from politicians. Actually so. And if you want to delay, she at at the time they were governors of different State services. So Joel dotted from so Paulo, and the directly from you 83 00:13:08.890 --> 00:13:11.349 Marina Fontolan: and both of them 84 00:13:11.590 --> 00:13:20.729 Marina Fontolan: for supporters of both sonaro in the 20 in the 2,018 elections. But after both sonaros 85 00:13:20.780 --> 00:13:26.619 Marina Fontolan: government, they of course, start to criticizing him, and they 86 00:13:27.100 --> 00:13:31.550 Marina Fontolan: so so 87 00:13:31.640 --> 00:13:41.119 Marina Fontolan: so they So they were into, of course, in supporting boson in 2,022. However, what happened during the elections was that 88 00:13:41.180 --> 00:14:10.690 Marina Fontolan: they so was that did this information pieces we're using both zone, the audience and the do. Our, the late just old videos to these inform people and tell them that they were supporting both sonaro in 2,022. So they take the video out of the context and reuse them in a new one The same thing happened with Andreja Noonis, who is a Federal Congressman nowadays, and he was 89 00:14:10.950 --> 00:14:21.200 Marina Fontolan: I mean he was. He became a very He became a Lula supporter, and he was the one that manage Lewis campaign 90 00:14:21.280 --> 00:14:31.390 Marina Fontolan: on social media during during the elections. And of course, and again, I mean, they use old videos often regenerate is criticizing Lula 91 00:14:31.430 --> 00:14:45.370 Marina Fontolan: to this in form to different people. and to state that they that he had to changes his heart or changes his mind. And now he was supporting both on. So in this case they were. They were like 92 00:14:45.630 --> 00:14:52.679 Marina Fontolan: reusing all the videos in a new context. So this was actually the best amount of the 93 00:14:53.140 --> 00:14:58.660 Marina Fontolan: the the. This was like the largest number of 94 00:14:58.960 --> 00:15:00.519 Marina Fontolan: of 95 00:15:00.910 --> 00:15:18.789 Marina Fontolan: of data that we have. So cases like that. However, it wasn't always It wasn't always these cases, And it wasn't always just about the politicians itself. 1 one example that you became very. 96 00:15:18.980 --> 00:15:34.670 Marina Fontolan: very famous here in Brazil. Was this video in which they recreated a news program here in Brazil that's called. Sure. Now, this is the most. Watch it. News watch news program here in Brazil 97 00:15:34.680 --> 00:15:36.160 Marina Fontolan: and 98 00:15:36.410 --> 00:15:55.510 Marina Fontolan: they created an entire deep fake video using Hernata Va vassel's voice to these days that there was a new poll a new elections full, and as you also now was the one that was that was the leader of the running up. 99 00:15:55.870 --> 00:16:14.889 Marina Fontolan: In the case of this video. What they did is that they switched the amount of supports from the polls between Bolsonaro and Booma. So Lula in this poll was had, like 40% of the intentions of the votes Bolsonar has around 30 100 00:16:14.900 --> 00:16:34.409 Marina Fontolan: But then in this video they swapped it. They said, they they made the entire video to make it look like and sound the like. Also saying that also not it was the running up with the 40%. And what was the second one with 30? So 101 00:16:34.520 --> 00:16:46.630 Marina Fontolan: they created this these it of of of like a parallel like a parallel universe. 102 00:16:46.660 --> 00:16:53.469 Marina Fontolan: for for these elections. So always building distrust on what was going on 103 00:16:53.810 --> 00:16:59.979 Marina Fontolan: The same thing happened during the World Cup. So this is a video of 104 00:17:00.030 --> 00:17:07.229 Marina Fontolan: the Amy is the the Emir from Katar. Tammin Hamas. I'll tell me. 105 00:17:07.490 --> 00:17:11.999 Marina Fontolan: I mean, of course. Qatar was hosting the World Cup at the time. 106 00:17:12.050 --> 00:17:26.569 Marina Fontolan: and of course, The Emir was talking about the word cup and etc. in his speech. But what they did is that they use the same video and they put new new subtitles, saying that 107 00:17:26.710 --> 00:17:33.880 Marina Fontolan: is that stated that the Qatar they put our in. Here was 108 00:17:34.000 --> 00:17:40.740 Marina Fontolan: was reporting that the Brazilian elections were frauded so. 109 00:17:40.920 --> 00:17:51.999 Marina Fontolan: which was a very sensitive situation situation here, and of course, I mean, as the vast majority of of Brazilians don't speak Arabic people 110 00:17:52.070 --> 00:17:54.829 Marina Fontolan: people would believe on the subtitles. 111 00:17:54.940 --> 00:18:02.540 Marina Fontolan: and also they did some like a very heavily editing on his Charleston's 112 00:18:02.590 --> 00:18:08.379 Marina Fontolan: speech saying, I mean he, Charleston, is a very famous 113 00:18:08.610 --> 00:18:14.589 Marina Fontolan: soccer player from the Brazilian team, and he was also 114 00:18:15.330 --> 00:18:29.670 Marina Fontolan: and the edited show and the edit and the editing made him sound like he was supporting both on and saying things like. Also that is the biggest Brazilian, the most patriotic one, and he's the one 115 00:18:29.780 --> 00:18:37.100 Marina Fontolan: and he's the one to be should be like a good president for the country. So 116 00:18:37.960 --> 00:18:45.020 Marina Fontolan: And of course I mean it wasn't just about soccer, which is very famous in Brazil, but also they used 117 00:18:45.120 --> 00:18:59.179 Marina Fontolan: They used a lot of editing for other situations. So this is Casino. He is a digital influencer here in Brazil. He do mostly react to videos, and he has, like millions and millions of followers. 118 00:18:59.260 --> 00:19:00.300 Marina Fontolan: So 119 00:19:00.620 --> 00:19:06.050 Marina Fontolan: at the time of the elections, he was celebrating his 20 ninth birthday. 120 00:19:06.130 --> 00:19:29.379 Marina Fontolan: But what they did is that they added edit his photo for his for his celebration, and put the number 22. So here in Brazil, our voting system is is, as I said electronically. it's the electronic key is everything's done electronically. And we and we have to type in numbers 121 00:19:29.420 --> 00:19:43.700 Marina Fontolan: to vote for a certain for to vote for for the people that we want to. So both on that was Number 22. So when they added the the photo, they took the 9 out, put the 22 and 122 00:19:43.800 --> 00:19:54.460 Marina Fontolan: and I started to share this image stating that was also supporting also, not because in you, of course, came to 123 00:19:54.910 --> 00:20:06.439 Marina Fontolan: came to public stating that he would never vote to both sonado and start to and start to publicly support Lula. So it was a very big backlash for the 124 00:20:06.560 --> 00:20:28.270 Marina Fontolan: for the for the for the government. And one thing that it's very important to also to mention here is, that's plavable. Sonaro is Jerry Boson out of son, and he's also a Congressman. So he was also involved in being like this information about spreading for his father. So that's 125 00:20:28.300 --> 00:20:32.190 Marina Fontolan: so. This also shows how 126 00:20:32.650 --> 00:20:46.380 Marina Fontolan: how they were operating at this time. And as a last example we had, like another video that was taken out of context also. and but in this case it was from variables on out of himself. 127 00:20:46.530 --> 00:20:53.800 Marina Fontolan: we chose that example of how the dis information may team 128 00:20:53.890 --> 00:21:05.709 Marina Fontolan: look like like a better like a good person and a good candidate and etc. So in this case, they, they took a video out of out of contacts in which 129 00:21:05.870 --> 00:21:25.270 Marina Fontolan: they said that Bolson was was speaking after the second turn of elections which he actually hadn't by the time, and he was in this case he was asking for like a military interaction. He would say that in this video that 130 00:21:25.420 --> 00:21:37.309 Marina Fontolan: the elections were frauded, and that the military should take actions here. here in Brazil. But again, it was something that he actually said, but not in this. 131 00:21:37.410 --> 00:21:44.210 Marina Fontolan: in this specific context of after after the results of the elections. 132 00:21:44.920 --> 00:21:50.979 Marina Fontolan: So one thing that it's that it's very 133 00:21:51.120 --> 00:22:14.220 Marina Fontolan: Important for us to think is that the expansion of production and sharing of fake and uses information, hate, speech, and intolerance brought to the process of social change that needs to be foundered and discussed with the participation of the several societal factors where it with the effects and effects of this phenomena in the democratic order. So 134 00:22:14.420 --> 00:22:25.860 Marina Fontolan: as as I was saying, and all this example shows what the I mean the they have like a clear vision of of watch. 135 00:22:25.930 --> 00:22:30.690 Marina Fontolan: of what to do with this information which was to build this trust in our 136 00:22:30.720 --> 00:22:41.590 Marina Fontolan: in our democracy, in our voting system, in our I mean in our elections, in the election process as a whole. So 137 00:22:41.750 --> 00:22:56.479 Marina Fontolan: the biggest result was actually a that actually failed. and it was carriage out on January 8. So it was like a week into Lula's government already. 138 00:22:56.540 --> 00:23:08.339 Marina Fontolan: and one thing and they and like both on our supporters, most of them were camping in front of the like, the military basis. 139 00:23:08.490 --> 00:23:18.730 Marina Fontolan: they They try to do like a cook. They they they destroyed the the buildings 140 00:23:18.810 --> 00:23:25.549 Marina Fontolan: several buildings from from Brazil that are famous that are used for G. French 141 00:23:26.080 --> 00:23:27.610 Marina Fontolan: for G, French 142 00:23:27.830 --> 00:23:37.319 Marina Fontolan: for G, for different powers here. So like our Congress, our Senate. and of course, I mean, did you? Did you just hear a system? 143 00:23:37.330 --> 00:24:03.389 Marina Fontolan: So they destroyed all these? all these buildings Many of them were arrested. So we had, like hundreds of people being arrested because of this, and the courage investigations of this group shows that the military were involved. A lot of business people, too. And also there is probably a lot of participation of also not and his family in this in this situation. So 144 00:24:03.470 --> 00:24:08.910 Marina Fontolan: so they used this information to build this trust on our system. 145 00:24:08.980 --> 00:24:11.810 Marina Fontolan: people reacted to it. 146 00:24:11.830 --> 00:24:21.839 Marina Fontolan: And this was the result. So just to finish through this information, the public sphere falls apart in this end of life. 147 00:24:22.170 --> 00:24:25.390 Marina Fontolan: So thank you very much from all of us. 148 00:24:27.560 --> 00:24:31.369 Thank you, Marina and Nicole, for this alerting presentation. 149 00:24:31.490 --> 00:24:43.610 And we now come from fake news to censorship, and I would like to ask Johann, if you guys know if it's pronounced 150 00:24:44.030 --> 00:24:46.600 a well to 151 00:24:46.790 --> 00:24:51.529 start with, of the presentation about the 152 00:24:52.160 --> 00:24:54.660 dual framework of saying censorship 153 00:24:56.620 --> 00:25:05.169 Zhifan Luo: thank you very much. Let me first try to share my screen and start my slides. 154 00:25:05.750 --> 00:25:12.869 Zhifan Luo: Hopefully, people can see me like see my slides now. But if not, please let me know. 155 00:25:12.870 --> 00:25:40.480 Zhifan Luo: So thank you so much to the organizers and to people who are attending around the world today, it's my great pleasure to present here. It's about 4 A. M. In Toronto right now, so I will try to keep away as much as I can, and hopefully I won't drop the board. In the next of 15 min or so I'm going to talk about the working paper, and I change the title a little bit, and the new title is censoring for legend, mercy, the social dimension of censorship. 156 00:25:40.830 --> 00:26:07.749 Zhifan Luo: When we talk about censorship, when we use the word censorship in the English language, we imply a system of political repression that restricts the freedom of speech, however, a censorship is not only just the political and politically imposed, but also socially constructed. for example, the Peo Research Center conducted a survey in the United States in 2,020, about people's attitudes towards cancer culture. 157 00:26:07.830 --> 00:26:14.599 Zhifan Luo: In that survey one question asks whether respondents saying a cancer culture is a form of censorship 158 00:26:14.700 --> 00:26:26.040 Zhifan Luo: about 26% of Republican respondents. Uncertainty. Yes, Council culture is censorship. Well, only about 6% of Democratic respondents answered, yes. 159 00:26:26.090 --> 00:26:32.140 Zhifan Luo: In other words, what is perceived as censorship is associated with people's political ideology. 160 00:26:32.550 --> 00:26:47.200 Zhifan Luo: for another example, about a month ago one of the major comedian company in China was fined for 2 million Us. Dollars, because one of its performers made a joke of the Chinese military in a live show. 161 00:26:47.570 --> 00:26:54.070 Zhifan Luo: What is not worthy in this event is that the joke was made in an offline live show. 162 00:26:54.140 --> 00:27:08.129 Zhifan Luo: So one of the audiences took the joke to the social media and the complain about it. It was only after that the official focused attention on this joke, and it took with the actions against the the performer and his company. 163 00:27:08.320 --> 00:27:24.639 Zhifan Luo: the first case, it's about censorship as the narrative and the perception and the segment case is about censorship as a way to control public speech. In both cases we cannot ignore the like. The social dimension of the role played by the public in it. 164 00:27:24.890 --> 00:27:39.679 Zhifan Luo: and, more interestingly, in the comedian taste, the public, the audiences made a more proactive role in censorship, that by helping the authority filter, what a speech is the judges, what speech is sensible! 165 00:27:40.060 --> 00:27:51.020 Zhifan Luo: The people who posted this joke, online, even publicly celebrated, is the fine of the comedian company as a victory. It indicates that a censorship may have public support. 166 00:27:51.590 --> 00:28:12.609 Zhifan Luo: This actually leads us to the key question of this study. The censorship started in the past is mostly built up on the assumption that a censorship as a form of political repression is against the the interest of the censored and the public. It is bad, it is yet legitimatable in the eyes of the censored and the public. 167 00:28:12.940 --> 00:28:16.819 Zhifan Luo: However, what if censorship is legitimate. 168 00:28:16.870 --> 00:28:30.899 Zhifan Luo: or whether censorship is legitimatable or not, should be started as an empirical question rather than a given Regardless of our researchers own personal opinion and moral conviction regarding censorship. 169 00:28:31.300 --> 00:28:47.500 Zhifan Luo: So in this case we can ask the questions about the legitimacy of censorship. How is this legitimacy produced, reproduced or challenged the in the public sphere, and they started actually fixing to the social expect of censorship the legitimacy 170 00:28:48.230 --> 00:28:56.640 Zhifan Luo: In the rest of my presentation, I will briefly introduce my theoretical framework on legitimacy and social construction of censorship. 171 00:28:56.650 --> 00:29:12.360 Zhifan Luo: Then I will introduce my case under study as well as my analytical strategies before. before I present some findings from the analysis, I will try to give some conclusions and touch on the implications at the end. If time permits 172 00:29:13.110 --> 00:29:17.899 Zhifan Luo: any, we can start with a quick review of the nature of sensorial power. 173 00:29:18.010 --> 00:29:41.079 Zhifan Luo: Well, most of the people think of censorship. We think of the We think of it like in terms of the liberal conception, as a form, as I said, as a form of political repression imposed by the authority upon Rose, who are censored. And this is the conception that you can find in most literature of censorship. 174 00:29:41.440 --> 00:30:01.270 Zhifan Luo: One of such alternatives is proposed as a so-called structured censorship on the basis of the theories by Michelle Focourt and Kirby. This series of power questions a distinction between external sensorial power and the autonomous individuals as assumed in the liberal conception. 175 00:30:01.420 --> 00:30:12.359 Zhifan Luo: In other words, censorship is not only measures taken to repress public speech, but it is also structural forces to shape the intent of expression at all. 176 00:30:12.540 --> 00:30:21.809 Zhifan Luo: so censorship can be structured and internalized. This strand of thinking is also labeled as the as a new censorship theory. 177 00:30:22.490 --> 00:30:37.229 Zhifan Luo: If we compare the liberal conception of censorship, and it's a new censorship theory. The key question at State is whether that's real power acts only repressively, or could they be employed in productive ways as well. 2. 178 00:30:37.990 --> 00:31:00.109 Zhifan Luo: What we can. I'm not getting into that debate right now, but we, what we can get from this debate from the new censorship theory is that a censorship should be understood not only in the political context that enables the States and other authorities to censor, but also in the social tongue heads that normalize, problematized and structuralized tensorial acts. 179 00:31:01.040 --> 00:31:06.000 Zhifan Luo: Indeed, scholars began to pay attention to social actors in censorship. 180 00:31:06.070 --> 00:31:22.369 Zhifan Luo: for example, in a study published the last year. we only, and I propose the concept of participatory censorship which focuses on the proactive participation in censoring public expression by ordinary social media users. 181 00:31:22.810 --> 00:31:31.060 Zhifan Luo: Also there is no secrets that pro-censorship attitudes exist in both authoritarian and the democratic societies. 182 00:31:31.820 --> 00:31:42.740 Zhifan Luo: whether it's social actors, rowing censorship or pro censorship attitudes. They demonstrate the diverse forms of censorship as social institutions. 183 00:31:42.880 --> 00:32:00.090 Zhifan Luo: and their effect as social institutions cannot be understood. If we merely took censorship as political institutions. and this is the reason why some censorship scholars call for a shift of tension from defining which forms of communication regulation count as censorship 184 00:32:00.310 --> 00:32:05.880 Zhifan Luo: towards defining how these different forms of censorship produce different effects. 185 00:32:06.850 --> 00:32:24.000 Zhifan Luo: and the to investigate an account for these such effects we need to understand as a moralization on censorship. For how censorship is moralized as a bad practice to repress and restrict, or as a good practice to protect and exert justice. 186 00:32:24.450 --> 00:32:37.230 Zhifan Luo: It's not only about whether public opinion supports censorship or not. It's also about the moralizing the public sphere and regulating what is acceptable and unacceptable collectively. 187 00:32:39.380 --> 00:32:55.429 Zhifan Luo: To demonstrate how social scientists can approach this broad issue. I start with the narrow question, key to the social dimension. The social institution of censorship is legit mercy. and like in the eyes of the subordinates. 188 00:32:55.569 --> 00:33:08.329 Zhifan Luo: I know legitimacy is another complicated term here. I will only focus on descriptive legitimacy in the with Arian sense. I'm not discussing prescriptive or nominative legitimacy 189 00:33:08.840 --> 00:33:29.460 Zhifan Luo: mit Ctl. And to empirically analyze descriptive legitimacy of censorship in the eyes of the subordinates I adopt the framework from David Bason. specifically, I analyze a legitimacy in terms of its justifiability through the sources of legitimacy, and as a content of established the rules one 190 00:33:29.590 --> 00:33:40.340 Zhifan Luo: also I analyze the consent to censorial act as voluntarily expressed through actions, and I will unpack this further in the findings. Part 191 00:33:41.300 --> 00:33:58.600 Zhifan Luo: for the empirical analysis, I conduct a case study of responsive censorship in China. In recent decades scholars have found that authoritarian regimes can be very responsive to the public complaints and the grievous. 192 00:33:58.630 --> 00:34:03.589 Zhifan Luo: So this idea of responsive authoritarianism has been proposed. 193 00:34:03.990 --> 00:34:12.430 Zhifan Luo: Here. I borrow this concept to name an act of censorship in response to public grievance of a drama show. 194 00:34:13.410 --> 00:34:28.160 Zhifan Luo: I keep the drama show anonymous to protect the social media users who commented on its being censored being taken down. But in general. It is an urban comedy show about the office romance. 195 00:34:28.590 --> 00:34:36.909 Zhifan Luo: This show has passed has been revealed by the censorial authority and the past, with a permission to air. Last year. 196 00:34:37.260 --> 00:34:50.959 Zhifan Luo: however, when it was aired on streaming websites, it actually arose the public range and a lot of debates on social media due to the vibe of the show, but also some very controversial things. 197 00:34:51.550 --> 00:35:03.059 Zhifan Luo: Here are 2 screenshots from the show in the left one the hand belongs to the male lead character, and the person in the we share is the female lead character. 198 00:35:03.230 --> 00:35:09.279 Zhifan Luo: In this screenshot you can see, the hand is pulling the stripe of the bra of the woman. 199 00:35:09.860 --> 00:35:15.409 Zhifan Luo: and in the right screenshot the men's hands is placed on the woman's breasts. 200 00:35:16.230 --> 00:35:21.339 Zhifan Luo: These 2 things are portrayed in non-critical way in the show. 201 00:35:21.490 --> 00:35:27.149 Zhifan Luo: It's just the to show like how these men making romantic advance towards the woman. 202 00:35:27.380 --> 00:35:34.849 Zhifan Luo: However, some audience did complain about it, because it looks more like sexual assault rather than romantic advance. 203 00:35:35.390 --> 00:35:43.700 Zhifan Luo: So the end of this controversy is that the show was up, rapidly taken down from all streaming websites on the same day. 204 00:35:43.870 --> 00:35:57.290 Zhifan Luo: which indicates that a higher authority above these streaming websites worse responsible for this coordinated pick down. which is also like the case of responsive censorship in my case. 205 00:35:58.490 --> 00:36:08.039 Zhifan Luo: with this case, I'm interested in 2 things. First, one is the take down of the show. The censorship is justifiable in China, in the Chinese context. 206 00:36:08.240 --> 00:36:19.850 Zhifan Luo: Second, yet the social media users voluntarily gave a consent to. They take down of the show. and this 2 interesting interest can be translated into 2 empirical questions. 207 00:36:19.860 --> 00:36:33.299 Zhifan Luo: First, it's the that real power and the play in this case justifiable, not second. How much did the public voluntarily express these? Then the consent or descent to the exercise of the all that the real power here 208 00:36:34.340 --> 00:36:46.760 Zhifan Luo: to answer. The first question I revealed the legal documents, including laws, regulations, departmental rules, and industrial regulations on censorship, of audio visual content in postmount China. 209 00:36:47.250 --> 00:37:02.930 Zhifan Luo: And to you answer the second question, I did a qualitative content. Analysis of all public posts commenting on the take down on Weibo, which is a major social media site in China, or you may think it. Think of it as the Chinese twitter 210 00:37:04.110 --> 00:37:22.700 Zhifan Luo: mit ctl. And in terms of the finding it's not difficult to find a legal justification of sensorial power. Here I present a sensorial class regarding gender and sex issues in audio visual products, in selective legal documents. And I just want to highlight 2 things, first, one 211 00:37:22.700 --> 00:37:34.580 Zhifan Luo: and 2 regulations in the 1990 s. The word should be removed from an audio video product includes any content of gender discrimination which basically can be applied to this case. 212 00:37:34.940 --> 00:37:41.350 Zhifan Luo: however, this clouds actually disappeared from legal documents since 1,997, 213 00:37:41.570 --> 00:38:01.699 Zhifan Luo: and at the same time more crowded censoring violations of traditional, sexual and marriage norms were added. Such content could include like extramarital sex and the romance, cohabitation, homosexuality. this kind of thing will be censored from audio visual content. 214 00:38:02.640 --> 00:38:21.420 Zhifan Luo: Eagle documents provide a basis of legality to the tensorial acts and power, and it is built upon the moral authority of the State to determine what is appropriate aid, gender, and a sex relationship, and what is inappropriate, and therefore has to be censored one. 215 00:38:21.720 --> 00:38:38.160 Zhifan Luo: However, this legality and the State's moral authority is actually in tension with each other. The Chinese day's moral authority used to be built up on its social to Transformation mission, a project of modernization and enlightenment. 216 00:38:38.260 --> 00:38:48.250 Zhifan Luo: However, it took a traditional turn, and the reason in that case to protect the more conventional forms of family and intimate relationships. 217 00:38:48.430 --> 00:38:56.770 Zhifan Luo: Therefore, censoring things of sexual assault or more traditional masculinity is at or with these recent laws. 218 00:38:56.820 --> 00:39:04.930 Zhifan Luo: In other words, it's not in the established return rules that the content producers will be healed. Accountable for this kind of patrol. 219 00:39:05.050 --> 00:39:18.529 Zhifan Luo: and, more importantly, it reveals that the controversy intrinsic to the State's moral authority, where it tries to protect conventional values while the citizens actually require otherwise. 220 00:39:19.350 --> 00:39:30.199 Zhifan Luo: So my next question is like, how did ordinary social media users think about this? Take them, do they, voluntarily consent to it? Or do they express descent? 221 00:39:30.560 --> 00:39:38.620 Zhifan Luo: A. And based on my content analysis, about a 45% of posts expressed the consent to the take that. 222 00:39:38.720 --> 00:39:45.089 Zhifan Luo: And these posts either praised this take down or celebrated as the late justice. 223 00:39:46.160 --> 00:40:01.800 Zhifan Luo: and about 21% of the posts explicitly expressed the descent to the pick down many of these polls that came from people who followed and watched the show. However, there are also arguments made on the rational legal ground. 224 00:40:01.920 --> 00:40:19.829 Zhifan Luo: There are people who didn't know what to show or like to show at, or but they were against these take down for it. It was arbitrary, because it was against the formal rules. Remember that the show has already passed the foremost sensor, real procedure to obtain the license. 225 00:40:20.010 --> 00:40:24.990 Zhifan Luo: So these people believe that rule is rule and the rule should be followed. 226 00:40:26.290 --> 00:40:37.889 Zhifan Luo: These are like 2 big stands regarding the take down. And behind these 2 stands I noticed that some main rationale of a particular interest to our understanding of legitimacy here. 227 00:40:37.950 --> 00:40:58.699 Zhifan Luo: The first one is mainly in the camp, or supporting the Ca, take down, of course, with some exceptions. the one focus the on the fact that is a show path, the formal content review, in the first place. But instead of insisting that a rule is rule. They ask the question, why? It wasn't a censored. In the first place, 250. 228 00:40:59.090 --> 00:41:05.539 Zhifan Luo: What's wrong with the sensor, real system. Why, it's not working, why it allows a show to be aired. 229 00:41:06.290 --> 00:41:10.220 Zhifan Luo: I summarize this as demand for better censorship 230 00:41:10.370 --> 00:41:23.309 Zhifan Luo: for a censorship that reflects the value system of the public which may not be aligned well with the conventional value system held up by the Chinese State, who took a traditional Tur. 231 00:41:23.680 --> 00:41:33.909 Zhifan Luo: So in this way censorship becomes a sort of public service, or what they people want is good public service that protect and reflect their values. 232 00:41:34.910 --> 00:41:41.139 Zhifan Luo: The last one is a relatively small group of posts are coming for about 15% of the data set. 233 00:41:41.220 --> 00:41:47.000 Zhifan Luo: They explicitly question the legitimacy of censorship, but on various grounds 234 00:41:47.320 --> 00:42:08.000 Zhifan Luo: in the first example. Here it questions the legit mercy on the rational legal grounds, about the arbitrariness of the take down, and in the stagnated tempo. It's again presental. Treat at the system, and it's that to be for the freedom of speech. But you can also see that this narrative is is inter coin within. In high families the narrative. 235 00:42:08.030 --> 00:42:23.469 Zhifan Luo: it regards families the ideology as the dominant one on the driving seat of the censorial authority, which is, of course, not true. It actually reminds me of the anti-censorship narrative regarding Kessel culture and a content moderation in the North American context. 236 00:42:24.290 --> 00:42:42.139 Zhifan Luo: So to sum up, censorship is an institution of political oppression, and I'm not questioning it here, but it is also a social institution, constantly produced and transformed through social processes and the interactions between the censors censored and the public one. 237 00:42:42.380 --> 00:43:00.050 Zhifan Luo: The key study of the real power in China shows that, like social scientists, can empirically analyze the censorship as a social institution. Here, since the real power targeting entertainment products can be justified on the basis of legality and the State's moral authority 238 00:43:00.260 --> 00:43:17.169 Zhifan Luo: and the voluntarily expressed the consent to the censorial action, also suggested some public consent to its legitimacy. However, the legal base of the tensorial power had its contradiction, and the public consent was not without a condition or a contention. 239 00:43:17.430 --> 00:43:28.860 Zhifan Luo: or though it's a real practice that enjoy the widespread support. In this case, I also observed a significant discontent. regarding the perceived, the failure to sensor properly. 240 00:43:29.220 --> 00:43:47.260 Zhifan Luo: and the sensor real authority received the criticism as their content. The review and a sense of real practice. They did not reflect the value system of the society. that's what I want to share, thanks for your attention, and I look forward to any questions or comments you may have. 241 00:43:51.560 --> 00:43:54.350 Yeah, thank you. She found for this. 242 00:43:54.400 --> 00:44:01.009 presentation about very intriguing topic. 243 00:44:01.620 --> 00:44:12.709 we have now plenty of time for questions and comments. So I ask everyone, either online or in this room to ask their question or 244 00:44:12.920 --> 00:44:21.069 restrict the comments? Yes. the first question in the room. Please come to the microphone or 245 00:44:21.280 --> 00:44:27.210 even here, you. So you can be seen by the people online. Yeah. 246 00:44:29.500 --> 00:44:33.859 And please, first tell your name and your affiliation. 247 00:44:34.280 --> 00:44:52.640 Hi! That's It's Tim Klaus. And so, from the Technical University of Boulogne I have a question to the last speaker to Louis Shivan, and thanks for an interesting talk, really interesting perspective on censorship. So my question is, I have 2 questions. Actually, one question is. 248 00:44:53.040 --> 00:45:11.399 are the Chinese authorities also use bots to influence the discourse about the take downs because you were speaking about very positive responses on way. Boo! That's saying always was great, that the TV show was, take down. And 249 00:45:11.440 --> 00:45:19.340 yeah, I'm just wanted to ask if they also use bots. And and the other thing is the other question. 250 00:45:19.680 --> 00:45:33.380 do the do the people that we port those? media that should be centered benefit in any way from it? What are they motivations? That's my question. Thank you. 251 00:45:33.960 --> 00:45:36.159 Shouldn't sit down. Yeah. 252 00:45:38.150 --> 00:45:41.630 you can answer directly. 253 00:45:42.020 --> 00:46:05.420 Zhifan Luo: yeah, sure. Thanks very much for the questions. for the first question, yeah, definitely, Chinese government is famous for using bugs and a buying bus, and sometimes their employees will do it themselves to like, try to influence the public opinion. And that's the reason why I choose this particular case. Because the content is 254 00:46:06.220 --> 00:46:09.749 Zhifan Luo: is like, it's nothing about politics. 255 00:46:09.890 --> 00:46:22.549 Zhifan Luo: It's just about office romance which a roots public rage somehow. So in my case, I would argue that the Government has some minimum interest 256 00:46:22.750 --> 00:46:26.359 Zhifan Luo: to influence public opinion, so 257 00:46:26.380 --> 00:47:02.910 Zhifan Luo: it does not apply to any case, but I would argue that in my case box is of that concern. But if I'm more talking about the comedian company and their jokes about the Chinese military. The case I mentioned at the beginning of my presentation I would be more worried about the Government trying to influence the public opinion and a shape of public opinion using bots and other forms. online to like, yeah, to shape the narrative so definitely, I think that's a very legitimate question. The second one, I think. 258 00:47:02.910 --> 00:47:07.149 Zhifan Luo: if I understand it correctly, you are asking about like. 259 00:47:07.330 --> 00:47:32.319 Zhifan Luo: what's the motivation for people to comments on this, take down right? Are there any commercial interest? Well, I would say in this particular case, because it happened very quickly, and I didn't show this bigger in this presentation. But actually, most the majority of the posts occurred. It was in the first one or 2 days online. 260 00:47:32.380 --> 00:47:58.959 Zhifan Luo: so I want to say that about that is very spontaneous in response to the sudden take down, because many people said about I'm just a watching the show, and I went to La to dinner when I came back that should already removed. So there is a lot of comments like this, so I would also argue that in this, in this case, it is things like, it's the stake is so low. And the people is just 261 00:47:59.590 --> 00:48:05.910 Zhifan Luo: expressing the opinion. But yeah, it's basically my argument. Thanks for your question. 262 00:48:08.360 --> 00:48:20.339 and that a direct question to this did they just express their opinion? Or did they say it's Such a show shouldn't be broadcast it 263 00:48:23.380 --> 00:48:25.699 Zhifan Luo: the show should be what 264 00:48:25.870 --> 00:48:27.750 it shouldn't be. Broadcast it. 265 00:48:28.020 --> 00:48:38.309 I want to say it. Did. They just express the opinion or they? They asked to put this down 266 00:48:39.230 --> 00:49:01.220 Zhifan Luo: So at the point of the my data collection of the show has already been removed. So they are basically just the celebrating, like in the first. first, the category people were just like, Wow, yeah. Finally, it should be removed a much earlier. But because I use the keywords the show has been taken down 267 00:49:01.230 --> 00:49:07.590 Zhifan Luo: to. So to like collect the data. So in my case, it's more about after the show has been removed. 268 00:49:08.010 --> 00:49:08.860 Okay. 269 00:49:12.350 --> 00:49:14.770 any other question or comment. 270 00:49:18.470 --> 00:49:22.139 Yeah, okay, please come here and use this microphone 271 00:49:22.270 --> 00:49:23.170 here 272 00:49:24.350 --> 00:49:40.249 and present yourself before asking, commenting. Okay, I use a few in a school of Solar University. Thank you very much for some interesting talk I want to ask. Shi funn. Also about this, now that you were on the topic like the 273 00:49:40.340 --> 00:49:51.569 the methodology behind. How you did this quality of content, math, because, as what I saw. It was only like 4 categories that you use, but it must have been some 274 00:49:51.740 --> 00:49:58.019 ambivalence or something landing like in between the categories. Or. 275 00:49:59.270 --> 00:50:01.510 yeah. thanks. 276 00:50:03.930 --> 00:50:28.909 Zhifan Luo: Yeah, definitely. So I cut off the methodology part from the presentation for the sake of the time. But yeah, definitely, master, it's it's very important here. So I want to explain my coding seem a little bit. I basically use emergent coding So I code a sample of the data set as a whole. That is about a 4 over 4,000. The way we will post the social media post. 277 00:50:28.910 --> 00:50:37.160 I selected a subset of from that to develop my coding thing and these coding categories. They are non exclusive. 278 00:50:38.070 --> 00:51:03.669 Zhifan Luo: So I code each category independently because you can be you. You might be get a surprise to how people can agree with the take a take down, and they disagree with it at the same time. So there's a lot of overlap and ambiguity here. that's the reason why I develop this coding thing, which includes 4 categories that are independent of each other. 279 00:51:03.890 --> 00:51:23.670 Zhifan Luo: So one poster can have a positive coding in the first. The category. It means it agree with the picked up, but also positive for the segment category, which means it also criticized the pick down somehow. yeah. So that's basically what I did here. 280 00:51:24.380 --> 00:51:38.410 and a quick, quick follow up. So you have coded the 5,000 posts. yes, I hand coded as a whole, that at that and myself. But I also have a research assistant, the coding, a sub-saharan 281 00:51:38.480 --> 00:51:49.070 Zhifan Luo: subset of it, just the for in your Coder reliability. When we reached a satisfactory intercoder reliability, we stopped there and I coded the rest of the data set. 282 00:51:49.140 --> 00:51:50.440 Okay, thank you. 283 00:51:51.710 --> 00:51:58.300 Thank you. We have another another question. Please come here and present yourself. 284 00:51:59.990 --> 00:52:08.109 Hello! My name is Nunavut Vida. I'm a researcher at Nova University of Lisbon, in Portugal. 285 00:52:08.430 --> 00:52:23.430 I have kind of a provocative question. If I put it at control it that way. It is very much inspired in the Balson Arrow case, but it can also be applied to to the other case or to any other cases. 286 00:52:23.470 --> 00:52:28.789 So I guess for me, the elephant in the room is. 287 00:52:29.040 --> 00:52:30.680 And 288 00:52:30.750 --> 00:52:35.840 yeah, whatever new kinds of technologies 289 00:52:35.900 --> 00:52:42.610 similar would appear. I'd like to understand your your point of view of 290 00:52:43.070 --> 00:52:51.850 of how the future will hold. I mean, when we look at the elections into Southern 22, or 23. 291 00:52:51.900 --> 00:53:00.570 Do you see yourself as this technology being advanced by either 292 00:53:01.250 --> 00:53:13.739 democratic countries or in non-so democratic or even detective ships, as being used to control and to 293 00:53:13.820 --> 00:53:17.699 manipulated the discourse in this in his cases. 294 00:53:17.740 --> 00:53:25.769 so for me, this is kind of a very big question, I know, but it's it's something that it's really in my mind. Thank you very much. 295 00:53:26.000 --> 00:53:28.150 I assume this question is 296 00:53:28.440 --> 00:53:43.649 Marina Fontolan: directed to Marina and Nicole, isn't it? Yeah, thanks. So, yeah, it's yeah. It's a big yeah, it's a big question. But I mean one thing that we. 297 00:53:44.650 --> 00:53:57.750 Marina Fontolan: I think that we should keep in mind is that like these things of these idea of like manipulation? it's already happening it. I think that, regardless of these in 298 00:53:57.800 --> 00:54:14.149 Marina Fontolan: this artificial intelligence technology. I mean, like social media. I mean, we do have like tons in terms of like literature. Now say, showing that the social media platforms they tend to go 299 00:54:14.170 --> 00:54:20.809 Marina Fontolan: they tend to like show more and more content that is like 300 00:54:20.970 --> 00:54:26.990 Marina Fontolan: radicalized and reduced to the right. So it's 301 00:54:27.230 --> 00:54:41.759 Marina Fontolan: So we do have already some some manipulation on this part. But of course, I mean, like artificial intelligence will be. I mean, it was already used in at least here in Brazil, in the 2020 302 00:54:41.850 --> 00:54:46.009 Marina Fontolan: case that I mean the one that I showed about 303 00:54:46.090 --> 00:55:01.579 Marina Fontolan: was, I mean, that was like everything was built through artificial intelligence. So even the speech of the news anchor and at the vessels was was created with with an AI bot 304 00:55:01.870 --> 00:55:16.360 Marina Fontolan: and then, of course, the the images and etc. So yeah, I mean, it's already happening. And of course, I mean, Rick. I think that's regulating social media would be like a good answer for us to 305 00:55:16.750 --> 00:55:21.479 Marina Fontolan: to start to start 306 00:55:22.860 --> 00:55:25.999 Marina Fontolan: for us to like. 307 00:55:26.040 --> 00:55:40.549 Marina Fontolan: Try to control a little of this the spread of this information, of course, there will always be this information of fortunately, and then, of course, regulating these artificial intelligence. software. And that's 308 00:55:40.760 --> 00:55:46.490 Marina Fontolan: that's how how I view things. I don't have any call once you to say anything else. 309 00:55:46.650 --> 00:55:50.699 Nicole De March: So about the technology we like, we are. 310 00:55:50.760 --> 00:56:04.140 Nicole De March: all the people like are really concerned so now with deep fake like Then you know of the truth, because people used to believe more and what they see 311 00:56:04.580 --> 00:56:09.500 Nicole De March: or then what they read. But the point is, 312 00:56:10.320 --> 00:56:11.890 Nicole De March: the chip fake 313 00:56:12.560 --> 00:56:34.560 Nicole De March: is like a we have to concern about the easy things you know, because of. the is a thing can be deceptive as a deep fake now. with the bomb bio phones. We can create things like it very easily. But we did even before then it was it to added. 314 00:56:34.570 --> 00:56:37.940 Nicole De March: and the thing works quickly, like in Brazil 315 00:56:38.060 --> 00:56:40.919 Nicole De March: and India. 316 00:56:41.210 --> 00:56:45.700 Nicole De March: they use a lot of like a we use in India a lot of Whatsapp. 317 00:56:46.110 --> 00:56:55.219 Nicole De March: So it's this, like a true spread this information with a image, because the person will not like analyze 318 00:56:55.850 --> 00:57:01.310 Nicole De March: each point. They? They just like a think, oh, this person 319 00:57:01.380 --> 00:57:02.720 Nicole De March: send you me 320 00:57:03.110 --> 00:57:15.640 Nicole De March: so I can trust this person, and they like it. Just send the emoji or the audio. The audio, like a is really concerning like you can mimic some voice. 321 00:57:15.660 --> 00:57:24.559 Nicole De March: And this is, you know, you don't need like a a deep fake to do that. And you just spread the thing. And a lot of people that like. 322 00:57:24.570 --> 00:57:27.239 Nicole De March: actually, the people are not educated 323 00:57:28.240 --> 00:57:31.140 Nicole De March: to check the information. 324 00:57:31.200 --> 00:57:39.890 Nicole De March: And was a lot of There was a lot of study in this information in text. 325 00:57:40.100 --> 00:57:42.010 Nicole De March: like, we know this 326 00:57:42.140 --> 00:57:55.030 Nicole De March: for a while, but the the problem is how this the fake checkers are going to analyze an image as all the other things. I like being more perfect 327 00:57:55.150 --> 00:58:03.560 Nicole De March: and trustable. But as I said, the problem when the the checkers like they show all this is fake 328 00:58:03.980 --> 00:58:08.400 Nicole De March: like a all. This information was a spread. So it's not quick 329 00:58:08.520 --> 00:58:10.030 Nicole De March: enough, you know. 330 00:58:11.000 --> 00:58:21.360 Nicole De March: to this. So the thing is, of course, like is debatable, and it seems like the this in for more. But we have 331 00:58:22.220 --> 00:58:34.139 Nicole De March: like? You remember that if you have a strategy, you can like a spread this information without easy thing that you can. Everyone can do, you know. 332 00:58:37.300 --> 00:58:39.210 Yeah, thank you. 333 00:58:39.700 --> 00:58:51.640 Maybe I've got another question for you, too. Marina Nicole. Have you done any network analysis how this 334 00:58:51.910 --> 00:58:53.720 fake news spread. 335 00:58:53.770 --> 00:58:58.740 and if there were some that were targeted to particular groups. 336 00:58:58.870 --> 00:59:00.130 or 337 00:59:03.020 --> 00:59:05.170 if they 338 00:59:06.350 --> 00:59:13.750 if they reach to a very large number of people on on both sides of the political spectrum. 339 00:59:14.280 --> 00:59:22.310 Marina Fontolan: Yeah. so these this, and this is actually being carried out in the, in the, in their project. 340 00:59:22.330 --> 00:59:30.870 Marina Fontolan: we are, we are analyzing Twitter data. So we are, we will be able to see exactly how 341 00:59:30.890 --> 00:59:39.289 Marina Fontolan: I mean how it went. But one thing that it's that it's for sure is that they spread was 342 00:59:39.480 --> 00:59:51.280 Marina Fontolan: was like big enough for this fact check agencies to sit down and actually and tech check the this information, because that's something that also happens 343 00:59:51.350 --> 00:59:57.529 Marina Fontolan: that. Also, I mean, it happens a lot to me. There is much more disinformation being created and shared online 344 00:59:57.560 --> 01:00:03.199 Marina Fontolan: then the fact checkers can can actually check. So 345 01:00:03.510 --> 01:00:15.580 Marina Fontolan: So I mean, if it got to the point that the fact checkers manage to shack it. Yeah, it means that it will that they consider to be like 346 01:00:15.870 --> 01:00:30.850 Marina Fontolan: far too spread to to not to be a concern. So so I mean, even though we don't have data yet to show on this. we are. I mean, we are sure that it was an it was something 347 01:00:30.870 --> 01:00:43.399 Marina Fontolan: we know that. I mean, considering how people reacted to to Mullah's to Mullah's election, I mean, people were camping in front of our military basis 348 01:00:43.650 --> 01:00:46.299 Marina Fontolan: here here in Brazil. 349 01:00:46.370 --> 01:00:53.760 Marina Fontolan: a lot of them a lot of them, I mean, all of them probably voted for 350 01:00:53.880 --> 01:01:09.779 Marina Fontolan: in the second in the second turn and they were and they were caps there waiting They were always saying to wait for the next 72 h, so like any, I think anyone would announce like a new 351 01:01:09.950 --> 01:01:21.520 Marina Fontolan: like a new thing, so the military would intervene, or people would have to take the matters on their own hands and etc. So that's exactly what happened on January eighth year. 352 01:01:21.660 --> 01:01:32.390 Marina Fontolan: So I don't think I am. It's probably probably some of these some of these. This information had had been spread 353 01:01:32.430 --> 01:01:45.589 Marina Fontolan: to other groups. So, for example, the the video of the Amir, the the mirrors talking about the present and the election probably went 354 01:01:45.760 --> 01:01:53.709 Marina Fontolan: which went viral to to like other bubble groups. Let's say it's not just the far right people. But 355 01:01:53.760 --> 01:02:00.479 Marina Fontolan: and we still don't have any data to show exactly how we went. So yeah. 356 01:02:00.940 --> 01:02:03.930 thank you. We have another question in the room. 357 01:02:10.550 --> 01:02:22.839 hey? It's me. And I have another question to Nicole and Marina. Thank you for the presentation, also the examples as a big soccer fan. I can also say that. 358 01:02:22.970 --> 01:02:52.679 Even in the German And they were overtaken. The information. That is, Charleston is for voting personnel. So that's very interesting. How the also media overtakes just those information. My question is, how is it? Distribution? between cheap fakes and deepfakes, because, as I understood it, you found more cheap fakes. 359 01:02:52.700 --> 01:02:58.369 then AI deepfakes and just wanted to hear something about that. Thanks. 360 01:02:59.040 --> 01:03:12.300 Marina Fontolan: Yeah, so yeah, I mean, in case of your, we actually, I think he actually supported most of that in the end. But that video specifically they, I mean he he hasn't said anything about 361 01:03:12.530 --> 01:03:20.550 Marina Fontolan: about those about the Bible sonar. So even though even though he hadn't like publicly said 362 01:03:20.590 --> 01:03:30.649 Marina Fontolan: anything about supporting also, not only the problem with I mean, at least on that video. He didn't say anything, but even though they did 363 01:03:30.920 --> 01:03:33.589 Marina Fontolan: so some of the 364 01:03:33.960 --> 01:03:35.570 Marina Fontolan: I mean they did edit 365 01:03:35.680 --> 01:03:45.739 Marina Fontolan: So yeah, we I mean, I don't know if Nicole once you talk more about the 366 01:03:45.830 --> 01:04:01.840 Nicole De March: sorry like to. But there was Charlotte. So in then he was, yeah, yeah, Neymar was a boson out of supporter because he didn't want to pay the text the text, yes. 367 01:04:01.870 --> 01:04:07.940 Nicole De March: so the he did a promise that if he he made like a a goal he's going like to. 368 01:04:08.220 --> 01:04:15.390 Nicole De March: She made the go to her, but he didn't. Yeah, because you, you, he can do that. 369 01:04:15.980 --> 01:04:22.499 Nicole De March: Yeah. But the thing, as I said, like there, there was more to it, because, like 370 01:04:22.570 --> 01:04:25.520 Nicole De March: it's kind of this quote 371 01:04:25.560 --> 01:04:41.539 Nicole De March: to do like a a good deep fake. But now it's it's becoming even easier, like I don't know if you already check like a Instagram or Tiktok, you can put like your face in a famous movie 372 01:04:41.620 --> 01:04:45.209 Nicole De March: like a big. You can use that. There are some apps 373 01:04:45.240 --> 01:04:51.020 Nicole De March: that you can do like a a deep sake in a real lifetime, you know. 374 01:04:51.140 --> 01:04:58.920 Nicole De March: like a you. I can do right now and use like a a, a phone hangs or something face, you know. 375 01:04:59.320 --> 01:05:02.579 Nicole De March: and something that is 376 01:05:02.740 --> 01:05:08.120 Nicole De March: There's not a lot of discuss. It is the like. 377 01:05:08.130 --> 01:05:13.809 Nicole De March: the synthetic voices like, I talked with 378 01:05:14.410 --> 01:05:25.720 Nicole De March: a researcher that research about that. And like he said that like we said, 10 min of my voice voice, or less he could like her, reproduce my voice. 379 01:05:26.090 --> 01:05:31.620 Nicole De March: But the thing is like you have to be more specialized to do that. But 380 01:05:33.230 --> 01:05:35.040 Nicole De March: it's this to deceive. 381 01:05:35.240 --> 01:05:42.260 Nicole De March: The thing is like. as Martin said, the people that are asking for intervention 382 01:05:42.680 --> 01:05:45.269 Nicole De March: they work it as a set. 383 01:05:45.410 --> 01:05:55.660 Nicole De March: but the difference, like when you think about the sap, if you think like they go to they I isolated in another geographical place. 384 01:05:55.980 --> 01:06:10.970 Nicole De March: But when you are in this group is that digital place, you know. So this information equates through, like the telegram Whatsapp. And so they, this, this information is spread 385 01:06:10.980 --> 01:06:12.510 Nicole De March: for like A, 386 01:06:13.500 --> 01:06:19.089 Nicole De March: they like the people that are already. I think they should trust in that. You know. 387 01:06:20.410 --> 01:06:24.319 Nicole De March: Sometimes people that are not like more neural are deceived. 388 01:06:25.100 --> 01:06:29.150 Nicole De March: but it initializes. You know 389 01:06:30.080 --> 01:06:39.020 Nicole De March: the manifestation that becomes the strongest. So we we saw that in 8 to January. How it works 390 01:06:41.130 --> 01:06:47.869 I don't know. If it answered the question. I don't know. It's like 391 01:06:49.010 --> 01:06:55.749 No other questions. 392 01:06:56.500 --> 01:06:58.510 Maybe I raise a 393 01:06:58.550 --> 01:07:05.860 a very silly question to all the 3 of us of you 394 01:07:06.420 --> 01:07:18.329 as censorship and fake news are threats to democracy. did you have any hints from your research 395 01:07:18.600 --> 01:07:25.080 what to do to contain these threats or to 396 01:07:27.740 --> 01:07:29.160 fight back? 397 01:07:31.650 --> 01:07:41.059 Marina Fontolan: Yeah. So like. at least in our in our case, here is like information, information, information, information. 398 01:07:41.130 --> 01:07:59.580 Marina Fontolan: we we need to regulate the the social media platforms regulate, actually regulate our digital lives. I mean, these big tech companies. at least here in Brazil they roll without let me. 399 01:07:59.730 --> 01:08:20.770 Marina Fontolan: They they? They do everything that they watch here they have. They haven't like any rules to follow. and here we are having some discussions about regulating regulating the the social media. And like the Internet Conference, so like these platforms will be held to be responsible for the call that that is shared 400 01:08:20.820 --> 01:08:30.599 Marina Fontolan: within within it. So they need to be more aware of what people are posting, what people are. 401 01:08:31.110 --> 01:08:35.920 Marina Fontolan: what people are sharing. I mean, because 402 01:08:36.510 --> 01:08:44.540 Marina Fontolan: because through these they are I mean, because, as as Nicole said to me, 403 01:08:44.560 --> 01:08:48.950 Marina Fontolan: the like, social media are creating this kind of cult. 404 01:08:49.140 --> 01:08:51.180 Marina Fontolan: that. 405 01:08:51.210 --> 01:08:53.750 Marina Fontolan: And people are starting to believe like. 406 01:08:54.170 --> 01:09:11.889 Marina Fontolan: And if they anything that it's posted on these like in close enclosed digital groups here. and I mean, it's not just here. We also saw in the United States, and the same thing happened there. So it's not, it's not something that we need to take lightly. 407 01:09:12.260 --> 01:09:18.470 Marina Fontolan: So yeah, at least, you're regulating regulating social media. That's that's for sure. 408 01:09:19.000 --> 01:09:39.549 So I didn't let me call or Vm, I guess. Here I he is someone who wants to answer to your position. hi again on the board of you there from Lisbon. I mean, this has been in my in my head for many months. And the easy answer is obviously regulate. 409 01:09:39.640 --> 01:09:52.190 And we regulate everything, and we think is going to actually have an effective role. But to be honest, blunt, and please forgive me 410 01:09:52.290 --> 01:09:54.800 to interrupt your train of thought, but 411 01:09:54.910 --> 01:10:01.979 we are beyond that. And that's why my first question actually addressed also authoritarian countries. 412 01:10:02.110 --> 01:10:11.990 Maybe in very good democratic countries, we can actually regulate plus educate. 413 01:10:12.240 --> 01:10:15.949 And by doing this we might achieve some sort of a 414 01:10:16.690 --> 01:10:26.319 a more consensual way of life, keeping democracy somehow alive. Maybe that's an hypothesis that I can conceive. 415 01:10:26.540 --> 01:10:30.700 But regarding the rest of the two-thirds of the world. 416 01:10:30.710 --> 01:10:36.140 I don't really think regulation plays any kind of role is just 417 01:10:36.320 --> 01:10:54.289 yeah. One easy answer to give. And I'm not just talking about China or whatever I mean if this is this goes much more beyond what we can control litigate. 418 01:10:54.460 --> 01:11:01.850 fine, whatever. This is going to be much deeper, and it's going to come as a tsunami. 419 01:11:02.080 --> 01:11:09.569 Marina Fontolan: That's why I was talking about the the elephant in the room. Is that okay? You did your studies in a moment 420 01:11:09.760 --> 01:11:11.399 in which okay. 421 01:11:11.610 --> 01:11:20.950 creating facts was kind of possible. But still we could actually analyze and have such some experts to say this was a fake. 422 01:11:20.970 --> 01:11:22.369 This was not. 423 01:11:22.400 --> 01:11:32.630 But from now onwards we chat, Tpt. And all these technologies this will be totally and controllable. So sorry to be 424 01:11:33.020 --> 01:11:52.340 Marina Fontolan: to be supplant. no, it's fine. I mean, like, I didn't. Yeah, I didn't talk about like education I completely. I actually completely forgot about it. Because currently here in Brazil, we are discussing like laws that are related to 425 01:11:52.420 --> 01:12:02.030 Marina Fontolan: to regulating social media here. That's why it came to my. It was the first thing that came to my to my mind. but I mean like. 426 01:12:02.360 --> 01:12:08.849 Marina Fontolan: when it comes to non-democratic countries, everything becomes 427 01:12:09.300 --> 01:12:11.930 Marina Fontolan: like moody or very 428 01:12:12.070 --> 01:12:29.989 Marina Fontolan: very easily. That's I mean, that's that's for sure. Every every time that I talk I talk from a like democratic perspective. So like I live in a democratic country. I mean, it's not this. Probably it's not like. It's not the longest lasting democracy 429 01:12:30.190 --> 01:12:37.519 Marina Fontolan: that we had. I mean, Brazil had. It's fair share of like dictatorships. throughout history. 430 01:12:37.540 --> 01:12:49.259 Marina Fontolan: but even though we are on my strong democracy now, and I and I spoke from this point from this point of view. So yeah, I mean it. It makes sense, you know your 431 01:12:49.460 --> 01:12:54.770 Marina Fontolan: your your point of view. I mean, that's something that we need to sit down and 432 01:12:54.940 --> 01:13:00.310 Marina Fontolan: and and probably try and and try to think about. 433 01:13:00.760 --> 01:13:08.080 Marina Fontolan: not just dictatorship, but also illegal Governments in this ideal, like governments did 434 01:13:09.090 --> 01:13:20.049 Marina Fontolan: that takes it themselves into being a democratic one. So so yeah, that's something that we need to. that's something that 435 01:13:20.300 --> 01:13:23.850 Marina Fontolan: our our research continue. We'll probably have to 436 01:13:24.060 --> 01:13:29.139 Marina Fontolan: to deal with. That's that's for sure I do with any call now or 437 01:13:29.650 --> 01:13:35.769 Marina Fontolan: no. The the only thing that I I was going to act like I agree with him, of course, like we have to 438 01:13:35.900 --> 01:13:39.930 Nicole De March: that have this program to educate people 439 01:13:40.150 --> 01:13:41.420 Nicole De March: to check 440 01:13:42.440 --> 01:13:47.189 Nicole De March: But at the same time. It's I was going to add about the 441 01:13:47.300 --> 01:13:52.529 Nicole De March: checks. It is because, like we, we are 442 01:13:52.600 --> 01:14:01.720 Nicole De March: our mainly focus. It was in auto-visual manipulation. like another other types of manipulation. But of course, like a 443 01:14:01.930 --> 01:14:12.230 Nicole De March: all the people were like scared with like the Pope imag it with the cold and everything like the the session Pope Image. 444 01:14:12.410 --> 01:14:15.609 Nicole De March: like in the thing, is becoming easier. 445 01:14:16.090 --> 01:14:21.250 Nicole De March: should do like there are a lot of arts and everything that you see like A, 446 01:14:21.570 --> 01:14:23.650 Nicole De March: yeah, they are using 447 01:14:24.580 --> 01:14:32.000 Nicole De March: like A, and we've chat deep. T, I, I think the other threat is like, it's easier 448 01:14:32.080 --> 01:14:34.790 Nicole De March: to program like to code. 449 01:14:35.470 --> 01:14:42.189 Nicole De March: And like I choose the cold to deceive, too, because then you can like it spread easily. 450 01:14:42.550 --> 01:14:50.770 Nicole De March: like a made some bots, I think, but I don't know in what stage of chats we are, you know. 451 01:14:50.940 --> 01:14:55.250 Nicole De March: So about about technology is the only thing that I I can. 452 01:14:55.700 --> 01:14:56.540 Hmm. 453 01:14:57.790 --> 01:14:58.900 Thank you. 454 01:14:58.990 --> 01:15:05.209 So I think with regard to censorship, and in China it 455 01:15:05.410 --> 01:15:08.050 might be a bit 456 01:15:08.070 --> 01:15:09.100 Tell me. 457 01:15:09.150 --> 01:15:20.340 uses to do something about it, because it's an authoritarian state. But you started when you started with cancel culture, also in democratic 458 01:15:20.460 --> 01:15:23.429 countries. And 459 01:15:23.830 --> 01:15:34.720 this kind of censorship from bottom up. Yet do you described them? Exist? Maybe also in this democratic countries. What? What do you think we we can do about it? 460 01:15:38.560 --> 01:15:56.099 Zhifan Luo: Yeah, definitely. I. I'm not sure whether it's my sociological training make me more cynical than I want myself to be. but just I then I also see the censorship, and sorry hearing contacts that can be coming from the button 461 01:15:56.100 --> 01:16:23.400 Zhifan Luo: just to add censorship, or I don't actually want to quit for council culture as censorship, but also like this kind of cancel culture coming from the bottom. So I would say, Yeah, the authoritarian context and the democratic context are extremely different in terms of the role played by the government and the regulations in the public sphere, but at at the same time the public's role in shaping the discourse, and in inter I 462 01:16:23.400 --> 01:16:41.569 Zhifan Luo: with this governmental forces. They have some similarity, and they can be compared to some extent. As I mentioned in my presentation, there is participatory censorship, which basically means, like the citizen in or coherent context that can use one. 463 01:16:41.580 --> 01:17:00.999 Zhifan Luo: the existing sense, a real regime for their own purpose. It's not for the government the purpose to spread a political ideology, but as a citizens can use the censorship regime to just just the to to do, to to achieve whatever goal they they want to achieve. 464 01:17:01.000 --> 01:17:11.600 Zhifan Luo: So in that way, I would say, even though it sounds very ironic. As a sociologist, I feels like it's an or about everyday resistance. 465 01:17:11.600 --> 01:17:13.439 Zhifan Luo: and every day 466 01:17:13.440 --> 01:17:32.830 Zhifan Luo: practice of the democratic power or the practice of trying to censor your peer citizen trying to strengthen the top down authoritarian control. So I would say it's actually up to everyone to resisting it. Otherwise, it's 467 01:17:33.150 --> 01:17:45.109 Zhifan Luo: because it's really hard to make sure policy works first. It's very hard to make a good policy. Second, is very hard to make sure good policy works. So 468 01:17:45.110 --> 01:18:08.789 Zhifan Luo: I'm not saying like policy does not matter. I do feel like it's very, very important, and it's have a huge effect. like the AI industry on how AI is a exercise the a. For example, in universities. Our university here in Canada are discussing our AI policy for the upcoming semester. So I would say, that is very important. 469 01:18:08.850 --> 01:18:26.780 But at the same time it feels very, also very important to focus on the everyday use every day interactions among citizens and civic groups. using the this technology using or this like institutional institutions. Yeah. 470 01:18:27.730 --> 01:18:38.899 yeah, thank you. I think this was a very good conclusion to our session. If there are any other questions here in the room. There are no questions among the people from the 471 01:18:39.020 --> 01:18:41.619 you are participating online. 472 01:18:42.190 --> 01:18:54.250 I don't see any questions, either, so I thank you all for your participation. I thank the 3 presenters for the very interesting and very initing presentations. And 473 01:18:54.390 --> 01:19:03.220 I wish you all and maybe Isabel has joined us. who wanted to moderate this 474 01:19:03.560 --> 01:19:10.129 session, but couldn't participate. So Is that what you like to say? Something in the end? 475 01:19:11.530 --> 01:19:34.559 Isabel da Costa: Hello, hello all. I'm very sorry I couldn't be there in person. We do. But I was following online. thank you all for your participation. Thank you so much, Stefan, for organizing this and sharing this, and you know that, together with the Marcus we're preparing a volume 476 01:19:34.600 --> 01:19:57.769 Isabel da Costa: with the the best presentation. So you've received. You've all received. communication about that. So if it's unpublished and you want to participate, please send this your communications your your proposals. Thank you, Stefan. It's been wonderful hearing, all of you. 477 01:19:57.770 --> 01:20:07.390 Isabel da Costa: Oops. Sorry and and I have a good continuation of the conference. Thank you all. 478 01:20:07.660 --> 01:20:10.310 Nicole De March: Thank you. 479 01:20:12.190 --> 01:20:13.190 Okay. 480 01:20:13.720 --> 01:20:25.130 so I wish everybody for of you a nice evening here in Australia, or a nice night in on other parts of the world, or a very nice day in 481 01:20:25.590 --> 01:20:30.060 another in in other parts of the world. So 482 01:20:30.360 --> 01:20:35.380 we are a global conference. And we have global continuing. 483 01:20:37.140 --> 01:20:38.579 Nicole De March: Thank you very much. 484 01:20:38.800 --> 01:20:41.000 Marina Fontolan: Thank you. Everybody. Thank you.