WEBVTT 00:00.000 --> 00:16.620 I'm going to introduce the next speaker, some of you that have been here since we've 00:16.620 --> 00:20.560 been going to give this a technical overview of where do you have to prepare them? 00:20.560 --> 00:23.460 This is Paralyze and there is Paralyze and there is Paralyze and there's a potential fact 00:23.460 --> 00:29.680 designing for trust and safety in the age of military technology, Paralyze and Mayze, and 00:29.680 --> 00:34.140 Mayze working with Paralyze, and here to space around the climate research is a fantastic 00:34.140 --> 00:39.680 architect, a little bit of a variety of sincerity at my arm, and I'm going to do a joint 00:39.680 --> 00:51.840 tool to work with it, take it away, let's try this one, yeah, I'm excited to be here, and 00:51.840 --> 00:55.160 I'm going to try another one, so they are going to talk about the signing for trust and safety, 00:55.160 --> 01:01.200 but specifically how designing for advocacy and security is a chain on my arm, and 01:01.200 --> 01:13.320 I'm going to say, I'm going to put out a communication and research and design, we are a 01:13.320 --> 01:19.260 think that we are facing the interconnects, we are ready to take the network, you know, 01:19.260 --> 01:37.120 we also do a lot of research, as well, into security design, a lot of the stuff we do, 01:37.120 --> 01:45.120 if you do a collaboration, yeah, always open for that, and so what I'm going to talk about, 01:45.120 --> 01:51.360 and I also need a resource design for it, is this important layer of design, so how product 01:51.360 --> 01:56.640 is design deeply matters for a year, can we do it to understand what I'm about to do, 01:56.640 --> 02:02.840 so we think of design and communication as a transition layer, can take what the policy of, let's 02:02.840 --> 02:07.840 say, if we can do that, the software, product, the technology, how it's built, and I really 02:08.080 --> 02:12.840 should translate it for facility, data experience, it's something that we just need to understand, 02:12.840 --> 02:17.800 the development of model around how that product functions, but design can be an unrivaled 02:17.800 --> 02:23.200 area, it can suppress content, it can let choose other content that's not about it, it's 02:23.200 --> 02:29.040 designed much like technology, it might be you, it can mislead you, and that's one of the 02:29.040 --> 02:35.840 things I also want to get into, and how that's like how design, the harmful design pattern 02:36.080 --> 02:44.080 it's designed to design patterns which are closely tailored and do it in fact, 02:45.960 --> 02:54.400 So I can even understand whether it's a bubler that has studied their technology so as to 02:54.400 --> 02:57.360 be good for modeling and understanding, the)] 03:01.060 --> 03:05.320 I'm going to get to trust as a user's cabinet, come into thisames round of which 03:05.320 --> 03:09.320 And it's also how much they can believe it's not a different one. 03:09.320 --> 03:11.320 This is very expensive. 03:11.320 --> 03:15.320 So you can be about reducing exposure escalation impact tomorrow. 03:15.320 --> 03:19.320 And so design where this later comes to the articulated. 03:19.320 --> 03:22.320 And again, this is our articulated two of these, right? 03:22.320 --> 03:27.320 Trust me, for the design or within the design player. 03:27.320 --> 03:29.320 I'm going to mislead you to see the design. 03:29.320 --> 03:33.320 And it's for filling on, like on this top. 03:33.320 --> 03:37.320 This design can also create your education. 03:37.320 --> 03:41.320 It's not particularly a professional student. 03:41.320 --> 03:43.320 It's not like I'm particularly designed. 03:43.320 --> 03:47.320 But I can't even do this with the importance of design. 03:47.320 --> 03:52.320 So then, if I wanted to talk about how that basically happened. 03:52.320 --> 03:55.320 So, for whole design patterns, often called the end of our pattern, 03:55.320 --> 03:57.320 except the design pattern. 03:57.320 --> 03:59.320 Our design patterns, that we understand, 03:59.320 --> 04:02.320 are potentially mislead to the community's knowledge, 04:02.320 --> 04:05.320 or how consumer-making decision, they normally look good. 04:05.320 --> 04:07.320 How many of you have tried that? 04:07.320 --> 04:09.320 I'm saying, scribe, we've got our product. 04:09.320 --> 04:11.320 We've got ourselves a scribe based later. 04:11.320 --> 04:13.320 We've got the idea of it. 04:13.320 --> 04:14.320 Yes. 04:14.320 --> 04:15.320 How do you get rid of it? 04:15.320 --> 04:18.320 I think you better to make the key to the arrow down. 04:18.320 --> 04:21.320 So, what's it to the trend of finding option at the moment? 04:21.320 --> 04:23.320 Very difficult. 04:23.320 --> 04:25.320 Either example to the point. 04:25.320 --> 04:29.320 But if you also include things like endless scroll, 04:29.320 --> 04:34.320 try to find a security tool or setting that's easy to get rid. 04:34.320 --> 04:35.320 Or according to the tools, 04:35.320 --> 04:37.320 if you will have to do a keychain, 04:37.320 --> 04:40.320 so if you're kind of a keychain on a platform, 04:40.320 --> 04:42.320 it's like someone's trying to contact you, 04:42.320 --> 04:43.320 but not yet. 04:43.320 --> 04:45.320 If you're also doing what you do, 04:45.320 --> 04:47.320 if you have to do the ecosystem, 04:47.320 --> 04:49.320 if you have a bit on it, 04:49.320 --> 04:51.320 then you can do it. 04:51.320 --> 04:54.320 So, these are all three different patterns. 04:55.320 --> 04:57.320 I've four of them, like hundreds of them, 04:57.320 --> 05:00.320 are placing the best intentions of the company. 05:00.320 --> 05:02.320 But I got two different things. 05:02.320 --> 05:05.320 So, here's the final one. 05:05.320 --> 05:07.320 I'm going to make quite a few. 05:07.320 --> 05:09.320 Is the training that you can see 05:09.320 --> 05:10.320 or is the training that you can see? 05:10.320 --> 05:12.320 Or is the training that you can see? 05:12.320 --> 05:15.320 Or is the training that you can see? 05:15.320 --> 05:17.320 No, I think this happens, 05:17.320 --> 05:19.320 but the ones you can see, 05:19.320 --> 05:21.320 they're not going to be like that. 05:22.320 --> 05:24.320 We're really going to have to do the training. 05:24.320 --> 05:26.320 And this is the framework of that. 05:26.320 --> 05:31.320 This is a thing that I would say 05:31.320 --> 05:33.320 in the presentation, which is a keychain, 05:33.320 --> 05:35.320 which is a keychain that you can see here. 05:35.320 --> 05:37.320 This is the keychain that you can see here. 05:37.320 --> 05:39.320 It's a free action to practice this into a 4K, 05:39.320 --> 05:41.320 which is a Dutch core exercise. 05:41.320 --> 05:42.320 It's a free-up. 05:42.320 --> 05:45.320 I think that is 12, which is a keychain. 05:45.320 --> 05:47.320 It's a one-cut-of-out. 05:47.320 --> 05:49.320 Well, we started looking at 05:50.320 --> 05:51.320 the recommended systems. 05:51.320 --> 05:53.320 Under the digital service tab, 05:53.320 --> 05:55.320 algorithmic, recommend your systems. 05:55.320 --> 05:58.320 Like, say, the algorithmic keychain 05:58.320 --> 05:59.320 is based on your Instagram. 05:59.320 --> 06:02.320 How to have a non-confident option. 06:02.320 --> 06:04.320 Most importantly, you should make that 06:04.320 --> 06:07.320 for a free period of chronological time. 06:07.320 --> 06:09.320 But it's hard to find 06:09.320 --> 06:11.320 but chronological time. 06:11.320 --> 06:13.320 Actually, I'm basically interested 06:13.320 --> 06:14.320 in this webinar. 06:14.320 --> 06:17.320 Our complaint was outlined in here. 06:17.320 --> 06:19.320 And so, it's algorithmic. 06:19.320 --> 06:21.320 It's algorithmic. 06:21.320 --> 06:23.320 It's very hard to find a line 06:23.320 --> 06:25.320 to go over the problem of the one. 06:25.320 --> 06:27.320 And so, we found that, notice that 06:27.320 --> 06:29.320 I highlighted a couple of articles with the DSA, 06:29.320 --> 06:31.320 because I also learned a whole bunch of patterns. 06:31.320 --> 06:34.320 So, the theory of the problem of the timeline 06:34.320 --> 06:36.320 is dark pattern. 06:36.320 --> 06:39.320 And then the ability to actually set 06:39.320 --> 06:40.320 to work on. 06:40.320 --> 06:42.320 Add it to the problem of the timeline. 06:42.320 --> 06:45.320 Also, by the way, let's do this. 06:46.320 --> 06:48.320 Everyone is a dark pattern. 06:48.320 --> 06:50.320 It has those D for some periods. 06:50.320 --> 06:52.320 Right, it's chronological D. 06:52.320 --> 06:54.320 It's a critical time. 06:54.320 --> 06:55.320 It's not set. 06:55.320 --> 06:58.320 It's a part of our research. 06:58.320 --> 07:02.320 These are showing me how to teach you 07:02.320 --> 07:04.320 that that's the algorithm of the timeline. 07:04.320 --> 07:07.320 So, that's the case with that. 07:07.320 --> 07:09.320 So, I just hope that this is fine. 07:09.320 --> 07:11.320 It's not easy to follow. 07:11.320 --> 07:13.320 It's easy to follow. 07:14.320 --> 07:17.320 I'm like, you see all the multiple steps. 07:17.320 --> 07:18.320 Right? 07:18.320 --> 07:19.320 This steps, this version, 07:19.320 --> 07:21.320 combined with the code's ball, 07:21.320 --> 07:23.320 which you know where to find it. 07:23.320 --> 07:24.320 Under this very setting, 07:24.320 --> 07:27.320 you know that you can set the dark pattern. 07:27.320 --> 07:30.320 And again, by the way, it's the TSA. 07:30.320 --> 07:32.320 Dark patterns can be traced 07:32.320 --> 07:34.320 with a more UX, at least issues. 07:34.320 --> 07:37.320 But you want to think of them as safety issues. 07:37.320 --> 07:40.320 Like, having a conflict that can play flow, 07:40.320 --> 07:42.320 vary, 07:42.320 --> 07:44.320 you know, tools that exist in a virtual reach, 07:44.320 --> 07:45.320 in a space. 07:45.320 --> 07:46.320 By the way, 07:46.320 --> 07:48.320 how many users should write 07:48.320 --> 07:50.320 so you can see what kind of artwork you also can see. 07:50.320 --> 07:51.320 Right? 07:51.320 --> 07:53.320 These are not just the method of the glitches, 07:53.320 --> 07:56.320 but the safety ones as well. 07:56.320 --> 07:57.320 So, 07:57.320 --> 07:59.320 this is where, like, 07:59.320 --> 08:00.320 these are difficult to get, 08:00.320 --> 08:01.320 and it is the design. 08:01.320 --> 08:03.320 Which is part of the design flow. 08:03.320 --> 08:04.320 Right? 08:04.320 --> 08:05.320 This is where it's flying. 08:05.320 --> 08:08.320 This is creating the space of the mark. 08:09.320 --> 08:10.320 So, 08:10.320 --> 08:12.320 I want to show that as an example, 08:12.320 --> 08:13.320 just to sort of, say, 08:13.320 --> 08:14.320 design, 08:14.320 --> 08:16.320 can have a sort of weaponized space. 08:16.320 --> 08:17.320 That's actually it. 08:17.320 --> 08:19.320 And in the research we do, 08:19.320 --> 08:22.320 we can be think of that this work would 08:22.320 --> 08:23.320 turn pattern in the circles. 08:23.320 --> 08:24.320 I said, 08:24.320 --> 08:25.320 when I said, 08:25.320 --> 08:26.320 when I said, 08:26.320 --> 08:27.320 when I said, 08:27.320 --> 08:28.320 when I said, 08:28.320 --> 08:29.320 when I said, 08:29.320 --> 08:30.320 when I said, 08:30.320 --> 08:31.320 when I said, 08:31.320 --> 08:32.320 when I said, 08:32.320 --> 08:33.320 when I said, 08:33.320 --> 08:34.320 when I said, 08:34.320 --> 08:35.320 when I said, 08:35.320 --> 08:36.320 when I said, 08:37.320 --> 08:39.320 when I said, 08:39.320 --> 08:40.320 I'm just saying, 08:40.320 --> 08:45.320 when I said. 08:45.320 --> 08:46.320 When I said, 08:46.320 --> 08:48.320 I'm just saying, 08:48.320 --> 08:49.320 you're not trusted, 08:49.320 --> 08:51.320 but I'm just saying, 08:51.320 --> 08:53.320 though it's true 08:53.320 --> 08:56.320 and it might be true probably, 08:56.320 --> 08:57.320 but it might be true it better's, 08:57.320 --> 08:58.320 if it is true and it may be true, 08:58.320 --> 09:00.320 and it may be true. 09:00.320 --> 09:02.320 and it may be true, 09:02.320 --> 09:05.320 and it may be because we are trained to do this thing, 09:05.320 --> 09:09.320 It's very, very smart, and that behavior is really, 09:09.320 --> 09:13.320 the ability to define those cool settings, 09:13.320 --> 09:16.320 but also the ability to do it, too. 09:16.320 --> 09:18.320 And the trusted takeover we've been doing, 09:18.320 --> 09:21.320 when you see privacy and security design 09:21.320 --> 09:23.320 well safety and proofs. 09:23.320 --> 09:25.320 And in order for folks to be targeted, 09:25.320 --> 09:27.320 if you have much more of a lot of settings 09:27.320 --> 09:29.320 for ground training than if you do it. 09:29.320 --> 09:32.320 You can make it easier for folks to report 09:32.320 --> 09:35.320 and escalate up use to content moderation, 09:35.320 --> 09:37.320 system to press and safety teams. 09:37.320 --> 09:39.320 We have more transparent systems, 09:39.320 --> 09:42.320 conspiracy, understand what the calorie is using. 09:42.320 --> 09:46.320 All of this really relates to the area of safety, 09:46.320 --> 09:47.320 per consumer. 09:47.320 --> 09:49.320 So we're able to see security come together 09:49.320 --> 09:53.320 to create an area where we can move the trust and safety. 09:53.320 --> 09:56.320 So spread it to the work that we're currently doing 09:56.320 --> 09:57.320 and just watching it. 09:57.320 --> 09:59.320 You might have been working with a German, 09:59.320 --> 10:01.320 and you don't want to keep moving. 10:01.320 --> 10:03.320 You can help each other to build out the definition 10:03.320 --> 10:06.320 of what we're going to see by design. 10:06.320 --> 10:08.320 You can approach them with their allergy. 10:08.320 --> 10:09.320 That's interesting. 10:09.320 --> 10:11.320 We're going to do some of those things. 10:11.320 --> 10:12.320 We're going to do some of those things. 10:12.320 --> 10:15.320 We're going to verify security by design. 10:15.320 --> 10:16.320 We're going to do some security by design. 10:16.320 --> 10:18.320 We're going to do some security in the beginning. 10:18.320 --> 10:20.320 And you're also thinking about it. 10:20.320 --> 10:21.320 All throughout the product. 10:21.320 --> 10:23.320 I'm going to send a link to our slides. 10:23.320 --> 10:25.320 Then I'm going to delete your centric. 10:25.320 --> 10:27.320 We're going to talk to them if you should provide it. 10:27.320 --> 10:29.320 You can see if you have to be the same thing. 10:29.320 --> 10:30.320 Thank you. 10:30.320 --> 10:31.320 Also, we're going to do all that. 10:31.320 --> 10:32.320 Very important. 10:32.320 --> 10:33.320 A lot of works. 10:33.320 --> 10:34.320 Been ahead of you. 10:34.320 --> 10:35.320 If you're centric. 10:35.320 --> 10:36.320 Safety. 10:36.320 --> 10:37.320 From all the way to the beginning. 10:37.320 --> 10:39.320 After saying as much. 10:39.320 --> 10:40.320 We're responding to it. 10:40.320 --> 10:41.320 You are. 10:41.320 --> 10:43.320 We'll see if we can do that. 10:47.320 --> 10:48.320 Yeah. 10:48.320 --> 10:49.320 So. 10:49.320 --> 10:50.320 Some of the things we've highlighted. 10:50.320 --> 10:51.320 I want to be right there. 10:51.320 --> 10:53.320 I'm not thinking of design. 10:53.320 --> 10:54.320 I'm not going to open it here. 10:54.320 --> 10:57.320 I'm going to come in to this part. 10:57.320 --> 10:58.320 So design. 10:58.320 --> 11:00.320 I'm going to find the technology. 11:00.320 --> 11:01.320 I'll see. 11:01.320 --> 11:02.320 We'll create. 11:02.320 --> 11:03.320 Safety. 11:03.320 --> 11:04.320 I'm not going to. 11:04.320 --> 11:05.320 Safety. 11:05.320 --> 11:06.320 So. 11:06.320 --> 11:07.320 So. 11:07.320 --> 11:08.320 We're going to. 11:08.320 --> 11:09.320 We're going to. 11:09.320 --> 11:10.320 So. 11:10.320 --> 11:11.320 So. 11:11.320 --> 11:12.320 So. 11:12.320 --> 11:13.320 So. 11:13.320 --> 11:14.320 So. 11:14.320 --> 11:15.320 So. 11:15.320 --> 11:16.320 So. 11:16.320 --> 11:17.320 So. 11:17.320 --> 11:18.320 So. 11:18.320 --> 11:19.320 So. 11:19.320 --> 11:20.320 So. 11:20.320 --> 11:21.320 So. 11:21.320 --> 11:22.320 So. 11:22.320 --> 11:32.320 So. 11:32.320 --> 11:37.320 So. 11:37.320 --> 11:38.320 So. 11:38.320 --> 11:39.320 So. 11:39.320 --> 11:41.320 So. 11:41.320 --> 11:42.320 Size. 11:42.320 --> 11:43.320 So. 11:43.320 --> 11:44.320 So. 11:44.320 --> 11:45.320 So. 11:45.320 --> 11:49.320 So. 11:49.320 --> 11:52.380 For the user events, he will have, as we work at it, 11:52.380 --> 11:54.560 just after we're going to pilot something. 11:54.560 --> 11:58.560 Here are some of the things that we researched yesterday. 11:58.560 --> 12:01.600 Or I'm going to be the easier to decide kind of like 12:01.600 --> 12:04.600 a Facebook list for example, if you can see what content 12:04.600 --> 12:07.040 you're going to be doing here, more of a Facebook 12:07.040 --> 12:08.720 or Skype, there's a lot of people who want some 12:08.720 --> 12:11.120 moderation in my opinion, but then what 12:11.120 --> 12:13.880 is all different types of social products, how 12:13.880 --> 12:18.920 that, for example, we have. 12:18.920 --> 12:22.800 So part of what we started to pull out 12:22.800 --> 12:24.680 was a project, what it's had. 12:24.680 --> 12:27.640 It's a group of elements, say, even more 12:27.640 --> 12:29.160 than I should do, right? 12:29.160 --> 12:32.040 Eight years, articles, and methodology that 12:32.040 --> 12:34.600 had existed, right now, I've been published over 12:34.600 --> 12:36.560 the past 20 years. 12:36.560 --> 12:39.520 So one of the useful things I'm going to be doing 12:39.520 --> 12:40.880 and how to be shown on a project, I 12:40.880 --> 12:43.120 call trust-and-triggering, between we 12:43.120 --> 12:46.200 were looking at, asking if everybody else were 12:46.200 --> 12:47.200 to be single. 12:47.200 --> 12:48.560 The answer is to realize that you're going to be 12:48.560 --> 12:52.400 able to be able to produce about 30 different countries. 12:52.400 --> 12:54.600 One thing that can out with this idea of social 12:54.600 --> 12:58.000 product, is that those two linear plates have been set. 12:58.000 --> 13:00.920 So, right now, if you finalize your report, 13:00.920 --> 13:03.000 you file a review for response. 13:03.000 --> 13:04.600 Why is it going to be more of a circular system 13:04.600 --> 13:05.960 that you have been shooting? 13:05.960 --> 13:08.320 We are going to draft a report or go get close. 13:08.320 --> 13:10.000 Characters of the plates, and bring that in, 13:10.000 --> 13:12.920 and then to another portion of the potting 13:12.920 --> 13:14.640 I do something at the end. 13:14.640 --> 13:17.040 Of all of the time, I don't know if someone 13:17.040 --> 13:19.040 is using a task to progress in the report, 13:19.040 --> 13:22.800 but I don't know if it's going to be able to do that. 13:22.800 --> 13:25.880 You're writing more than a solution for an operation. 13:25.880 --> 13:28.640 So you need to break out of the server, or hold it. 13:28.640 --> 13:32.040 Okay, do you find a specific person? 13:32.040 --> 13:35.240 Pretty much as well, I'm asking the issue of tools. 13:35.240 --> 13:36.640 And if they think of the strategy, 13:36.640 --> 13:39.080 like this piece of consent, and more consent, 13:39.080 --> 13:42.960 the process of reporting, so like, ensuring that, 13:42.960 --> 13:45.720 no, it's one rather than what we've said. 13:45.720 --> 13:47.280 We've got it, but if you're a similar one, 13:47.280 --> 13:48.880 it's not rejected. 13:48.880 --> 13:51.760 We'll talk about it in a minute. 13:51.760 --> 13:53.320 Oh, is it not working? 13:53.320 --> 13:56.800 I don't know, I don't know, I don't know. 13:56.800 --> 13:58.400 I don't know. 13:58.400 --> 13:59.120 I know, you don't know. 13:59.120 --> 14:02.560 And that, oh, it hasn't up in recording this whole time. 14:05.760 --> 14:08.160 I guess a whipping fee a technology conference 14:08.160 --> 14:11.760 without a technology issue. 14:11.760 --> 14:15.320 Dang, that's okay, it's fine. 14:15.320 --> 14:18.080 Anyway, I'll just keep going. 14:18.080 --> 14:22.000 Anywho, you all look a little low energy 14:22.000 --> 14:24.200 because it's such an uplifting topic 14:24.200 --> 14:28.640 that I'm talking about, or is it because it's hot in here? 14:28.640 --> 14:30.600 Maybe it's both. 14:30.600 --> 14:34.040 Anywho, I guess, jumping back into this. 14:34.040 --> 14:35.640 For designers and for companies, 14:35.640 --> 14:37.240 one of the also big recommendations 14:37.240 --> 14:39.760 is can you take trauma-informed design principles? 14:39.760 --> 14:40.880 If you're wondering what those are, 14:40.880 --> 14:44.400 we are going to discuss them on the next slide, 14:44.400 --> 14:47.160 conducting threat modeling during development phases. 14:47.160 --> 14:48.600 Is this something from our research? 14:48.600 --> 14:50.720 We also saw that a lot of trust and safety teams 14:50.720 --> 14:52.680 either didn't know what that was, 14:52.680 --> 14:54.160 or not actively conducting. 14:54.160 --> 14:57.480 They were sort of designing for a handful of personas 14:57.480 --> 15:01.280 versus sort of taking a much more expansive view. 15:01.280 --> 15:03.440 And a big one, prioritizing user safety 15:03.440 --> 15:05.560 over engagement metrics. 15:05.560 --> 15:08.160 So again, really thinking about that some folks might be 15:08.160 --> 15:10.280 posting content that they don't want a lot of people 15:10.280 --> 15:12.400 to look at, and that's OK, and that's their choice, 15:12.400 --> 15:14.840 and that's also a safety choice. 15:14.840 --> 15:17.280 And including marginalized communities 15:17.280 --> 15:21.880 in the design process, one of the other elements 15:21.880 --> 15:25.160 of this research is chains, trauma-informed design principles 15:25.160 --> 15:28.520 where a big fan of them, they're a UK-based NGO. 15:28.520 --> 15:32.440 And so their trauma-informed design principles 15:32.440 --> 15:35.040 are foundational to everything they do in build. 15:35.040 --> 15:37.040 They also build software. 15:37.040 --> 15:41.800 And so for them, this is where they're experiencing, 15:41.800 --> 15:43.320 they're building around the experiences 15:43.320 --> 15:45.320 of survivors of gender-based violence. 15:45.320 --> 15:48.040 That also includes folks facing intimate partner 15:48.040 --> 15:50.720 violence, for example. 15:50.720 --> 15:54.600 For them, and this is all read from their principles. 15:54.600 --> 15:56.040 So this is what they've written. 15:56.040 --> 15:58.160 This is not my writing. 15:58.160 --> 15:59.880 So they write quote. 15:59.880 --> 16:01.880 Survivors are not just users. 16:01.880 --> 16:03.720 They are people navigating trauma. 16:03.720 --> 16:05.480 They're people in survival mode. 16:05.520 --> 16:07.800 So keep us in mind, you're designing a, 16:07.800 --> 16:09.800 you're when you're designing that service. 16:09.800 --> 16:12.440 If not designing with survivors in mind, 16:12.440 --> 16:14.360 it could be triggering trauma-in-packs 16:14.360 --> 16:16.320 all areas of grain. 16:16.320 --> 16:17.840 So what does this mean in practice? 16:17.840 --> 16:20.600 So safety, quick exit buttons, 16:20.600 --> 16:23.040 avoid overloading survivors, 16:23.040 --> 16:26.440 simple designs to ensure emotional safety 16:26.440 --> 16:27.880 and lessen fatigue. 16:27.880 --> 16:31.640 Fewer clicks, meaning like do not bury info. 16:31.640 --> 16:33.600 Privacy is non-negotiable. 16:33.600 --> 16:37.480 So data minimization, the ability to delete 16:37.480 --> 16:39.200 an account and data at any time, 16:39.200 --> 16:40.920 and that has to be prioritized 16:40.920 --> 16:43.120 from the outside of building, right? 16:43.120 --> 16:45.920 So again, easy deletion, accountability. 16:45.920 --> 16:48.800 So explaining our decisions, transparency, 16:48.800 --> 16:51.840 open and publicly, trust is built this way. 16:51.840 --> 16:55.640 Constantly measuring the impact when you release something, 16:55.640 --> 16:57.240 so how you know it's going to work 16:57.240 --> 16:58.720 and how many mechanisms. 16:58.720 --> 17:01.080 And I think within that, they also go further 17:01.080 --> 17:03.440 and describe impact for, again, 17:03.440 --> 17:06.200 very specific impacted groups, not just every user, 17:06.200 --> 17:08.440 is at working also with the most vulnerable 17:08.440 --> 17:10.920 or the most marginalized. 17:10.920 --> 17:13.680 And I also want to sort of share some work 17:13.680 --> 17:16.240 from this other group called PIN America. 17:16.240 --> 17:17.880 Again, they work a lot with journalists. 17:17.880 --> 17:19.920 And so there's three reports we looked at. 17:19.920 --> 17:22.160 One was called the OXQS for Abuse. 17:22.160 --> 17:24.320 And within OXQS for Abuse, 17:24.320 --> 17:27.120 there's some of the specific things they found 17:27.120 --> 17:31.680 where, so they're trying to encourage users rather 17:31.680 --> 17:34.200 than just punish abusers. 17:34.200 --> 17:35.960 Some of the specific design suggestions 17:35.960 --> 17:39.040 they found from interviewing human rights experts, 17:39.040 --> 17:40.480 designers, technologists, and journalists, 17:40.480 --> 17:43.880 are things like a shield and dashboard system. 17:43.880 --> 17:46.440 So a way to quarantine a piece of content 17:46.440 --> 17:48.560 for safety review, safety mode, 17:48.560 --> 17:50.800 so like a one-click thing where you can change 17:50.800 --> 17:52.560 all of your settings quickly if you're facing, 17:52.560 --> 17:54.800 let's say like dog pilot, a lot of people 17:54.800 --> 17:57.720 are jumping on your account. 17:57.720 --> 17:59.880 Rapid response teams have trusted allies 17:59.880 --> 18:02.480 to help that person facing harassment or harm, 18:02.480 --> 18:05.400 manage that abuse, and automatic documentation tools 18:05.400 --> 18:06.840 for legal evidence. 18:06.840 --> 18:09.560 So if someone has sent you like a series of threats, 18:09.560 --> 18:11.360 being able to quickly document that 18:11.360 --> 18:15.160 from the platform itself, transparency, penalty systems 18:15.160 --> 18:17.120 with escalating consequences. 18:17.120 --> 18:18.400 One of the things that their research 18:18.400 --> 18:21.200 and our research has found is that on the other side, 18:21.200 --> 18:24.760 if you are an in person, engaging in abuse, 18:24.760 --> 18:27.120 but more generally, no user actually knows 18:27.120 --> 18:30.160 like where they sit, how many strikes are against them 18:30.160 --> 18:33.200 in a platform, there are strikes against you. 18:33.200 --> 18:35.760 And so you can be like walked out of your account, 18:35.760 --> 18:38.920 let's say, for a day or a week, 18:38.920 --> 18:41.920 and you might not know why, and that could be from like, 18:41.920 --> 18:43.800 historical strikes that have been against you. 18:43.800 --> 18:46.920 So this does, this is something that's a big issue 18:46.920 --> 18:50.440 on all sides of people engaging online, 18:50.440 --> 18:52.640 but yeah, this transparent penalty system. 18:52.640 --> 18:54.400 And again, a trauma informed reporting 18:54.400 --> 18:56.320 for coordinated attacks, 18:56.320 --> 18:58.960 maybe an easy, easy to report. 19:01.200 --> 19:02.200 But yeah. 19:02.200 --> 19:06.840 And then one of the other things we also looked at 19:06.840 --> 19:11.840 was this worker by the National Democratic Institute. 19:12.560 --> 19:16.000 So this is a tracker, which is a structured data set 19:16.000 --> 19:18.480 of interventions, studies, and recommendations related 19:18.480 --> 19:19.840 to online gender based violence 19:19.840 --> 19:23.880 of abuse, disinformation, and other forms of online harm. 19:23.880 --> 19:26.800 When we were developing the safety by design framework, 19:26.800 --> 19:29.160 in which we're launching in this merch, 19:29.160 --> 19:32.520 the NDI landscape tracker really helped us identify 19:32.520 --> 19:34.160 reoccurring themes and interventions 19:34.160 --> 19:36.520 by systematically reviewing the interventions 19:36.520 --> 19:37.240 in the tracker. 19:37.240 --> 19:39.840 So one of the things we're publishing is a large data set, 19:39.840 --> 19:42.440 which will also be able to access, which is, 19:42.440 --> 19:45.240 pretty much every provocation and proposal 19:45.240 --> 19:47.440 of a safety intervention and policy 19:47.440 --> 19:51.560 we could find published through papers online. 19:51.560 --> 19:53.120 So again, these are the things we're talking about 19:53.120 --> 19:55.800 for folks like what a folks historically requested. 19:55.800 --> 19:58.440 And we're trying to tag as best we can, 19:58.440 --> 20:00.360 what's been implemented. 20:00.360 --> 20:02.120 Is it perfect? 20:02.120 --> 20:06.400 What we're doing now, is it incomplete, probably. 20:06.400 --> 20:08.120 But it's built off some of this hard work 20:08.120 --> 20:10.640 that NDI has done in the past. 20:12.280 --> 20:13.520 One of the things we're able to see 20:13.520 --> 20:16.360 is what we're calling common categories of action. 20:16.360 --> 20:19.800 And those are examples of platform transparency, 20:19.800 --> 20:21.000 reporting mechanisms. 20:21.000 --> 20:22.560 So the data set we're putting together 20:22.560 --> 20:25.560 and that NDI is put together goes into much more depth 20:25.560 --> 20:28.800 around an example of a reporting mechanism. 20:29.920 --> 20:32.640 Legislative and policy recommendations, 20:32.640 --> 20:36.000 training and capacity building examples, 20:36.000 --> 20:38.000 research and visibility of harm patterns, 20:38.000 --> 20:42.040 including hyperspecific patterns of harm 20:42.040 --> 20:45.200 and harassment, and localized or context-sensitive 20:45.200 --> 20:46.560 protective measures. 20:47.840 --> 20:50.720 We in our data set, 20:50.800 --> 20:53.440 view these reoccurring themes that can be translated 20:53.440 --> 20:55.760 into design and interventions to reference 20:55.760 --> 20:59.480 and integrate into findings and recommendations. 20:59.480 --> 21:03.800 So this tracker in particular aggregates actual documented 21:03.800 --> 21:05.400 interventions and research findings 21:05.400 --> 21:08.080 from multiple organizations and context. 21:08.080 --> 21:09.640 This isn't an incredibly large spreadsheet 21:09.640 --> 21:10.960 that we're all so linking to. 21:10.960 --> 21:13.040 So if you'd like access to this, 21:13.040 --> 21:15.160 let me know if it is publicly accessible, 21:15.160 --> 21:17.080 I can share it out later. 21:17.080 --> 21:22.080 Okay, best practices, I'm almost done. 21:22.440 --> 21:25.600 So in this last section, 21:25.600 --> 21:29.040 I want to shift our focus to also what can actually be done. 21:29.040 --> 21:30.600 Arriving, we've discussed so far 21:30.600 --> 21:33.600 from harmful design patterns to safety by design. 21:33.600 --> 21:36.280 And the frameworks and examples serve all point. 21:36.280 --> 21:40.520 We think to this simple and heavy quotes conclusion. 21:40.520 --> 21:43.560 So safety cannot be under tackled by one team 21:43.560 --> 21:46.680 or one discipline as requires multi-stake holder 21:46.680 --> 21:49.360 responsibility and action across a product. 21:49.360 --> 21:52.600 So it can't just be isolated to one trust and safety team, 21:52.600 --> 21:54.600 multiple teams are the entire product 21:54.600 --> 21:57.600 have to also be dedicated to centering safety. 21:58.800 --> 22:00.520 And I think this is true if you're trying 22:00.520 --> 22:02.440 to build a security culture as well. 22:02.440 --> 22:05.840 You have all the teams have to be dedicated to this. 22:08.520 --> 22:09.840 There you go. 22:09.840 --> 22:14.840 So what does this serve multi-stake holder responsibility mean? 22:14.840 --> 22:17.600 Design must be usable, accessible, and legibles. 22:17.600 --> 22:18.920 It has to be easy to find. 22:18.920 --> 22:22.320 Again, building on some of the points that Elio and Ania 22:22.320 --> 22:24.080 outlined in their talk. 22:24.080 --> 22:26.360 Design and technology have to be trauma informed. 22:26.360 --> 22:28.280 That does mean thinking about how much information 22:28.280 --> 22:29.800 or you're giving people, is it something 22:29.800 --> 22:31.760 that in a moment of panic? 22:31.760 --> 22:33.480 Can people understand? 22:33.480 --> 22:34.320 Can they access? 22:34.320 --> 22:35.320 Can they sift through? 22:35.320 --> 22:36.960 Is it something that will be triggering to them? 22:36.960 --> 22:38.520 Is it something that, again, they can, like, 22:38.520 --> 22:42.480 they have the cognitive bandwidth to engage with? 22:42.480 --> 22:45.840 If it's multiple lines of small text, the answer is no. 22:45.840 --> 22:48.800 That's very hard to read for people there in moments 22:48.800 --> 22:51.320 of high panic or high stress, right? 22:53.520 --> 22:55.720 A big thing, which I'm sure all of you do in this room, 22:55.720 --> 22:57.960 but the big tech does not. 22:57.960 --> 23:00.080 Centering opt-out over-opt-it. 23:00.080 --> 23:04.320 That means making sure all the privacy settings are turned on. 23:04.320 --> 23:06.640 Versus having to go and find them 23:06.640 --> 23:10.800 and make your account more and more private, right? 23:10.800 --> 23:15.000 This ensures security and safety are already 23:15.000 --> 23:17.040 set for the most safe user experience 23:17.040 --> 23:19.120 and folks can go toggle as they feel comfortable 23:19.120 --> 23:22.040 for how public they would like to be or how exposed, 23:22.040 --> 23:24.760 they'd like their data to be. 23:24.760 --> 23:28.560 Design should take into the emotional state of the user. 23:28.560 --> 23:31.240 So a big one, as we've been saying earlier, 23:31.240 --> 23:34.240 design was repeatedly center the most marginalized 23:34.240 --> 23:37.360 or as some folks say the decentered user. 23:37.360 --> 23:39.120 Design should actively threat model 23:39.120 --> 23:43.000 and understand and prepare for all these nuanced types of harm. 23:43.000 --> 23:47.840 Design is not neutral, so every default, every flow, 23:47.840 --> 23:50.320 every hitting, setting, encodes assumptions 23:50.320 --> 23:52.880 about power, risk, and responsibility. 23:52.880 --> 23:55.480 These are things we have to actively tackle 23:55.480 --> 23:59.720 in a design practice and also within product design. 23:59.720 --> 24:03.840 So I just want to end on this quote that we love. 24:03.840 --> 24:06.240 We use a lot, this is from off-sonnerry got, 24:06.240 --> 24:09.480 the creator of the decentered and from this paper 24:09.480 --> 24:12.120 she wrote called design from the margins. 24:12.120 --> 24:15.400 So she says, we expect architects to design buildings 24:15.400 --> 24:19.040 and British to withstand jail, force wins, and heavy loads. 24:19.040 --> 24:22.560 So two should we expect companies whose products are the vehicle 24:22.560 --> 24:25.640 for free expression and access to information 24:25.640 --> 24:27.920 to design sensitive, resilient technologies 24:27.920 --> 24:30.520 for the benefit of all, whether they face government 24:30.520 --> 24:33.760 or pressure and social stigma or not. 24:34.400 --> 24:35.560 Thank you. 24:35.560 --> 24:47.040 If you have any question or access to any of the materials 24:47.040 --> 24:48.440 I've shared today, please email me. 24:48.440 --> 24:51.880 I would happily share them out with all of you. 24:51.880 --> 24:53.840 So thank you. 24:53.840 --> 24:55.320 Let's keep that up for a second. 24:55.320 --> 24:57.480 I think we have two more minutes, but... 24:57.480 --> 24:59.000 Oh, shit, I'm sorry. 24:59.000 --> 25:04.000 But I think that's an off-sonner for a simple, 25:04.000 --> 25:07.800 one of the best ones, yeah. 25:07.800 --> 25:09.560 Yeah, I can hear you. 25:09.560 --> 25:10.560 Yeah. 25:10.560 --> 25:28.560 So that sort of relates to like, when you're at the same boat. 25:29.560 --> 25:32.280 Sorry, that's the first and asking a question. 25:32.280 --> 25:35.680 My friends was asking, how can design take into the emotional 25:35.680 --> 25:37.440 state of the user how would you know? 25:37.440 --> 25:39.760 So when you're thinking about some of these broader principles 25:39.760 --> 25:44.160 you're talking about like legibility or accessibility, right? 25:44.160 --> 25:45.960 I think those go hand in hand with this. 25:45.960 --> 25:49.120 Like is the text small, or is it easier to read? 25:49.120 --> 25:54.120 Is it using like jargon, or is it using plain language? 25:54.120 --> 25:58.360 Is it, is it a lot of text someone has to stitch through? 25:58.360 --> 26:02.560 So making something that has a larger font is using more plain language 26:02.560 --> 26:05.400 is easier to find easier to surface. 26:05.400 --> 26:06.760 Like those are already things that are 26:06.760 --> 26:09.440 to take into the emotional state of a user. 26:09.440 --> 26:11.360 Generally, if you're designing, let's say, 26:11.360 --> 26:14.720 for accessibility and you're designing for a wide audience, 26:14.720 --> 26:17.640 you're probably already centering. 26:17.640 --> 26:20.320 The emotional state is you're making it easier 26:20.320 --> 26:22.640 to find and surface that information 26:22.640 --> 26:26.720 and you're really taking into account how people 26:26.720 --> 26:30.280 are understanding or could misunderstand what it is you're outlining. 26:30.280 --> 26:34.640 But a big thing is making sure you're not giving them a lot of small text. 26:34.640 --> 26:38.600 Even when they get to a page, let's say of resources, right? 26:38.600 --> 26:43.480 So even if they can find your resource, making sure that it's not like a wall 26:43.480 --> 26:46.240 of multiple paragraphs of stuff they need to do. 26:46.240 --> 26:49.320 So if you are building, let's say, like an anti-doxing tool 26:49.320 --> 26:51.760 and you're trying to help them remove information, 26:51.760 --> 26:54.640 it may be telling them, hey, it's going to take this many minutes 26:54.640 --> 26:58.360 to go through this and you're really trying to space out some of the information. 26:58.360 --> 27:05.640 You're also making sure it's this balance right of your 27:05.640 --> 27:07.120 ideally, hopefully hiring a copywriter. 27:07.120 --> 27:11.640 But you're really trying to make sure the information is bite size 27:11.640 --> 27:13.200 as well as relevant. 27:13.200 --> 27:15.800 So it's like kind of cutting out extraneous information. 27:15.800 --> 27:30.800 Oh wait, sorry. 27:30.800 --> 27:31.800 Oh yeah. 27:31.800 --> 27:32.800 Oh, we did get certain. 27:32.800 --> 27:33.800 It's okay. 27:33.800 --> 27:34.800 Yeah. 27:34.800 --> 27:44.800 So sometimes not though. 27:44.800 --> 27:45.800 Yeah. 27:45.800 --> 27:48.800 How do you see design? 27:48.800 --> 27:50.800 Safety by design. 27:50.800 --> 27:51.800 Yeah. 27:51.800 --> 27:52.800 That's a great question. 27:52.800 --> 27:56.160 So I mean, one of the reasons I will never say, oh, yeah. 27:56.160 --> 27:59.800 So this person has a question of, you know, that harmful design patterns 27:59.800 --> 28:03.800 can are intentional by platforms or companies. 28:03.800 --> 28:05.800 And how do we see safety by designing being adopted? 28:05.800 --> 28:09.800 I will never say in a recorded session that they're intentional 28:10.800 --> 28:14.800 because I have to work on cases where it's impossible to prove intentionality. 28:14.800 --> 28:19.800 As an FYI, but we can have a solid conversation about that. 28:19.800 --> 28:23.800 But one of the things we believe at convocation is that when you design 28:23.800 --> 28:28.800 for a world that creates incentives, safety, security, and privacy, 28:28.800 --> 28:33.800 you would actually mitigate a lot of dark patterns or harmful design patterns 28:33.800 --> 28:37.800 because you're creating a world that centers the user, their agency 28:37.800 --> 28:38.800 and effectively their choice. 28:38.800 --> 28:43.800 So you're making it easier for them to really go through and make decisions. 28:43.800 --> 28:48.800 If you are in a world where you're trying to center privacy and use your agency, 28:48.800 --> 28:52.800 you are thinking about how to, I think, allegedly share that content 28:52.800 --> 28:55.800 and not judge a user towards something more harmful. 28:55.800 --> 29:00.800 Generally harmful design patterns are prioritizing a company and its interests 29:00.800 --> 29:03.800 over consumers and generally their safety. 29:03.800 --> 29:07.800 I like making our data as public as possible. 29:07.800 --> 29:10.800 So I do think that these go hand in hand. 29:10.800 --> 29:14.800 It does mean a lot of things would be like for radically redesigned. 29:14.800 --> 29:17.800 So if someone wants to unsubscribe, that'd be a very easy flow. 29:17.800 --> 29:20.800 There's probably a lot of reasons they want to unsubscribe. 29:20.800 --> 29:24.800 But it might be like a longer ripple effect of safety by design. 29:24.800 --> 29:28.800 We don't mention that explicitly in our paper, the unsubscribe example. 29:28.800 --> 29:33.800 But we do mention how easy some of these safety features have to be to find. 29:33.800 --> 29:37.800 And so that would be like pro user or user-centered settings, 29:37.800 --> 29:39.800 as perhaps a way to think about it. 29:39.800 --> 29:42.800 So they do go, I think, hand or at least they do, 29:42.800 --> 29:47.800 with how we think about approaching design and technology policy. 29:47.800 --> 29:49.800 Do you think you're in a question? 29:49.800 --> 29:51.800 Do you have a normal? 29:51.800 --> 29:52.800 Oh, yeah. 29:52.800 --> 29:57.800 So a lot from making the companies the course that we've been doing behind us 29:57.800 --> 30:00.800 and the designers, you're not to buy organizations. 30:00.800 --> 30:03.800 Most of the things that you would be up to, 30:03.800 --> 30:05.800 you wouldn't have an opportunity to decide 30:05.800 --> 30:07.800 if it works some reason or not for another. 30:07.800 --> 30:08.800 Yeah, thank you for that. 30:08.800 --> 30:12.800 So the question was, I guess, outside of trying to force companies with law 30:12.800 --> 30:14.800 or legislation to implement some of these principles 30:14.800 --> 30:18.800 are their other ways to encourage them or have them implement them. 30:18.800 --> 30:21.800 I mean, I would say so this might sound naive, 30:21.800 --> 30:25.800 but I, again, have worked at Wikimedia on the Trust and Safety team 30:25.800 --> 30:29.800 and that I think is a wonderful project where we center 30:29.800 --> 30:31.800 or try to center as best we could. 30:31.800 --> 30:34.800 They're like the requests and wishes of users. 30:34.800 --> 30:38.800 I do think when you treat users as real people 30:38.800 --> 30:42.800 who make your product what it is and you're trying to center 30:42.800 --> 30:46.800 and you're working with them and trying to center what they want 30:46.800 --> 30:49.800 and how they want to sort of exist and navigate in a project 30:49.800 --> 30:52.800 including your safety, you are taking them seriously 30:52.800 --> 30:54.800 and what you do is you end up building trust with them. 30:54.800 --> 30:58.800 And I think I think trust is an impossible thing to purchase. 30:58.800 --> 31:02.800 It's a very difficult thing to build and it's a very easy thing to lose, 31:02.800 --> 31:06.800 but it is what keeps consumers and users returning. 31:06.800 --> 31:09.800 It's what builds a community into a real community. 31:09.800 --> 31:11.800 So I think I recently said as naive, 31:11.800 --> 31:15.800 I think that that's a reason to design and to build that into product design 31:15.800 --> 31:19.800 because I think that that's an important important thing for user retention. 31:20.800 --> 31:25.800 But I don't work at a big tech company so I'm sure they have responses 31:25.800 --> 31:27.800 as to how that's naive. 31:27.800 --> 31:29.800 But I don't think that it is. 31:29.800 --> 31:30.800 Yeah. 31:30.800 --> 31:31.800 You can tell.