[00:01.120 --> 00:03.020] Oh, I guess it is. [00:03.820 --> 00:05.180] I guess it is. [00:05.180 --> 00:07.060] There we go. Awesome. [00:07.060 --> 00:09.500] Because as we were just sitting here joking, [00:09.500 --> 00:11.260] we're like, uh, this is absolutely [00:11.260 --> 00:13.060] the slot to see who [00:13.060 --> 00:14.660] made it to bed early. [00:14.660 --> 00:17.020] And who has already, ooh, [00:17.020 --> 00:18.600] and now I'm everywhere. [00:18.700 --> 00:20.500] This is awesome. [00:20.500 --> 00:23.360] We called it the hangover slot. [00:23.360 --> 00:24.660] So welcome. [00:24.880 --> 00:27.080] Yeah. And of course we only [00:27.080 --> 00:29.040] have one microphone, so the good [00:29.040 --> 00:31.240] news is that works better. [00:31.320 --> 00:33.300] So welcome, everyone. [00:33.640 --> 00:35.100] And we're going to take a little [00:35.100 --> 00:37.000] bit of a different approach, [00:37.000 --> 00:39.200] I guess, than just presentation [00:39.200 --> 00:40.640] slash panel. [00:40.780 --> 00:42.780] Because one of the [00:42.780 --> 00:44.900] issues with this is [00:44.900 --> 00:46.940] it's an evolving space when we're [00:46.940 --> 00:48.760] talking about the data and the privacy [00:48.760 --> 00:51.020] and all the issues with, [00:51.020 --> 00:53.240] I mean, if you ask anyone to define [00:53.240 --> 00:55.480] as Britton's going to talk about, [00:55.480 --> 00:57.400] like, some of the aspects [00:57.400 --> 00:59.540] of the data privacy in XR. [00:59.540 --> 01:01.700] I mean, we don't even have a common definition [01:01.700 --> 01:03.660] for some of the technology, [01:03.660 --> 01:05.620] so then how are you going to define what it's [01:05.620 --> 01:07.960] doing? So I'm Liz, [01:07.960 --> 01:09.680] and this is Britton, and [01:09.680 --> 01:11.660] we're going to talk, first Britton's [01:11.660 --> 01:13.480] going to lay a great foundation [01:13.480 --> 01:15.300] and give us a framework [01:15.300 --> 01:17.280] from which to have a discussion. [01:17.300 --> 01:19.620] And then the idea is we'll talk [01:19.620 --> 01:21.460] amongst ourselves, and then once y'all have [01:21.460 --> 01:22.660] had an opportunity [01:24.360 --> 01:26.480] to wake up, [01:26.480 --> 01:29.300] get the brain cells firing, [01:29.300 --> 01:30.600] then we thought we would open [01:30.600 --> 01:32.520] it up to a conversation and [01:32.520 --> 01:34.780] questions from the group. [01:34.780 --> 01:36.420] So, bear with us, [01:36.420 --> 01:38.600] and I'm going to let a much smarter person [01:38.600 --> 01:40.520] than I introduce the [01:40.520 --> 01:41.620] topic and [01:42.420 --> 01:44.480] explain a little bit of where some of the [01:44.480 --> 01:45.720] challenges lie. [01:48.740 --> 01:51.320] Question one, can you hear me? [01:51.600 --> 01:52.560] Awesome. [01:52.620 --> 01:54.600] Question two, who's heard [01:54.600 --> 01:56.080] of Lego Universe? [01:57.540 --> 01:58.620] Alright, [01:58.620 --> 02:00.420] I have a salty story [02:00.420 --> 02:02.300] for your Sunday morning about [02:02.300 --> 02:04.390] Lego Universe. [02:05.120 --> 02:06.320] Lego Universe [02:06.320 --> 02:08.100] was a highly [02:08.100 --> 02:10.220] anticipated massive multiplayer [02:10.220 --> 02:12.600] online game, like Roblox, [02:12.600 --> 02:14.360] Minecraft, that [02:14.360 --> 02:16.250] ended up being shuttered. [02:16.960 --> 02:18.320] The experience looked [02:18.320 --> 02:20.280] great, and the company [02:20.280 --> 02:22.320] was really looking to allow [02:22.320 --> 02:24.480] users to build their own type [02:24.480 --> 02:27.100] of content and their own environments. [02:27.400 --> 02:28.540] But given how utterly [02:28.540 --> 02:30.840] trusted Lego was as a [02:30.840 --> 02:32.700] brand, they committed to [02:33.500 --> 02:35.580] whitelisting content before [02:35.580 --> 02:37.800] it went live. So they were going to preview [02:37.800 --> 02:39.340] everything to make sure it was safe [02:39.340 --> 02:42.460] for you and your kids. [02:42.460 --> 02:43.620] The only problem was that [02:43.620 --> 02:47.240] there's a term in content [02:47.240 --> 02:47.700] moderation that [02:47.700 --> 02:49.460] Lego did not think about, and [02:49.460 --> 02:52.240] it's called time to penis. [02:50.480 --> 02:52.480] Time to penis [02:52.240 --> 02:53.860] literally means [02:53.860 --> 02:56.100] the duration [02:56.100 --> 02:58.040] between a person entering a virtual [02:58.040 --> 03:00.260] world and adult content [03:00.260 --> 03:03.400] appearing in that virtual space. [03:03.500 --> 03:04.280] And with all puns [03:04.280 --> 03:06.120] intended for Lego Universe, [03:06.120 --> 03:08.560] the time to penis was not very long. [03:10.520 --> 03:12.280] Lego needed, amongst [03:12.280 --> 03:13.660] other things, to [03:14.240 --> 03:16.920] figure out an automated, [03:16.920 --> 03:18.500] scalable, economical [03:18.500 --> 03:20.520] way to prevent teenagers [03:20.520 --> 03:22.840] and kids from drawing penises [03:22.840 --> 03:24.580] in the virtual world, [03:24.580 --> 03:26.480] and they couldn't do it. [03:26.740 --> 03:28.560] It became a wide-scale [03:28.560 --> 03:30.760] problem, and it actually ended up killing [03:30.760 --> 03:32.200] the project. [03:33.220 --> 03:34.600] This is not an [03:34.600 --> 03:36.780] unusual story. Many [03:36.780 --> 03:38.520] companies find that their expectations [03:38.520 --> 03:40.700] about governing and moderating [03:40.700 --> 03:42.460] user behavior and [03:42.460 --> 03:44.560] user data flows and spatial [03:44.560 --> 03:47.440] computing are turned on their head. [03:48.440 --> 03:48.740] My [03:48.740 --> 03:50.560] name is Britton Heller, and I [03:50.560 --> 03:52.600] like to describe myself as the lawyer [03:52.600 --> 03:54.120] of the black mirror. [03:54.760 --> 03:56.680] I am a visiting scholar at the [03:56.680 --> 03:58.600] Virtual Human Interaction Lab at Stanford [03:58.600 --> 04:00.500] University, and I teach [04:00.500 --> 04:03.060] international law and AI policy. [04:03.060 --> 04:04.820] And I've been focusing on immersive [04:04.820 --> 04:07.060] environments for about seven years. [04:08.260 --> 04:10.700] There are three things that I would like you to [04:10.700 --> 04:12.060] take away today. [04:12.400 --> 04:14.680] One, immersive media [04:14.680 --> 04:16.840] has a different impact on our [04:16.840 --> 04:18.970] bodies and our minds. [04:19.680 --> 04:20.880] And this is what [04:20.880 --> 04:22.980] differentiates it from flat-screen [04:22.980 --> 04:24.700] social media. [04:24.920 --> 04:26.680] Existing privacy law [04:26.680 --> 04:28.860] does not accommodate for [04:28.860 --> 04:30.080] this. [04:30.600 --> 04:32.600] Two, the different technical [04:32.600 --> 04:35.160] stacks between flat-screen [04:35.160 --> 04:37.000] social media and [04:37.000 --> 04:38.940] immersive content [04:38.940 --> 04:40.840] mean that we have different [04:40.840 --> 04:43.020] layers of enforcement when you're [04:43.020 --> 04:45.120] looking at behavioral tracking [04:45.120 --> 04:47.260] and modification in XR. [04:48.020 --> 04:49.180] And three, AI [04:49.180 --> 04:51.260] will not save us. And that one is [04:51.260 --> 04:52.920] worth repeating twice. AI [04:52.920 --> 04:55.300] will not save us. [04:55.640 --> 04:57.200] Overall, the systems that we [04:57.200 --> 04:58.800] create in 3D environments [04:58.800 --> 05:01.020] are fundamentally different [05:01.020 --> 05:02.780] from 2D environments [05:02.780 --> 05:05.020] because of the technical stack, the [05:05.020 --> 05:07.000] potential impact on users, and the [05:07.000 --> 05:09.280] uninformed state of the law. [05:09.720 --> 05:11.020] This leads us to three [05:11.020 --> 05:13.080] questions. What kinds of data flows [05:13.080 --> 05:15.520] do you get from XR environments? [05:15.520 --> 05:16.920] What does this mean for business [05:16.920 --> 05:19.000] models and the law? And why does [05:19.000 --> 05:20.500] this matter to users? [05:23.180 --> 05:24.780] Imagine that you and I [05:24.780 --> 05:27.160] are playing a race [05:27.160 --> 05:29.560] car-oriented game in XR. [05:30.200 --> 05:31.180] And I see this [05:31.180 --> 05:33.180] red McLaren, and I [05:33.180 --> 05:35.160] really, really like this car. [05:35.160 --> 05:37.120] I was going to say a Tesla, but the one that they're [05:37.120 --> 05:39.260] hacking over there, I don't know if it's actually going to be able to drive [05:39.260 --> 05:41.280] after this week. So, I [05:41.280 --> 05:43.020] like the Tesla. I like [05:43.020 --> 05:44.760] McLaren. You can like the Tesla. [05:46.340 --> 05:47.220] When I see [05:47.220 --> 05:49.360] this car, my body starts [05:49.360 --> 05:51.220] to physiologically react [05:51.220 --> 05:53.060] to my pleasure. So, my [05:53.060 --> 05:55.160] heart rate goes up a little bit. [05:55.160 --> 05:57.260] My skin gets a little more moist. [05:57.260 --> 05:59.060] My pupils dilate. [05:59.180 --> 06:00.860] I really like this car. [06:01.860 --> 06:03.180] Later on, as we're walking around [06:03.180 --> 06:06.060] the conference, and we go into a different experience, [06:06.060 --> 06:07.260] I see [06:07.640 --> 06:09.160] a similar red car drive [06:09.160 --> 06:11.300] by, driven by someone who looks [06:11.300 --> 06:13.180] more than a little bit like me. [06:13.980 --> 06:15.200] I go to a virtual sports [06:15.200 --> 06:17.580] bar, and someone who looks exactly [06:17.580 --> 06:19.060] like my crush [06:19.060 --> 06:21.160] says, isn't it a great [06:21.160 --> 06:23.340] time to buy a new car right now? [06:23.800 --> 06:25.200] I go to my email, and [06:25.200 --> 06:27.320] I start getting targeted ads from [06:27.320 --> 06:29.260] Geico about why I should [06:29.260 --> 06:31.180] re-up my car insurance. [06:31.980 --> 06:33.240] This sounds like [06:33.240 --> 06:35.520] science fiction, but the framework [06:35.520 --> 06:37.180] for it actually already [06:37.180 --> 06:39.180] exists. Companies have been [06:39.180 --> 06:41.640] putting advertisements in XR environments [06:41.640 --> 06:42.460] since [06:43.940 --> 06:45.440] 2021, I believe [06:45.440 --> 06:47.200] it was announced. And I've [06:47.200 --> 06:49.240] written papers about how advertisements [06:49.240 --> 06:51.200] in immersive contexts are [06:51.200 --> 06:53.180] actually experiential. [06:53.400 --> 06:55.240] They don't look like billboards. [06:55.240 --> 06:57.320] You actually pay for the privilege of having [06:57.320 --> 06:59.400] the car advertised to you. [07:00.380 --> 07:01.440] I developed [07:01.640 --> 07:02.740] a term called [07:02.740 --> 07:04.480] biometric psychography [07:04.480 --> 07:06.140] to describe [07:06.140 --> 07:08.920] the potential for this harm. [07:09.300 --> 07:10.240] Where your [07:10.240 --> 07:12.540] likes, dislikes, and consumer [07:12.540 --> 07:14.580] preferences are married [07:14.580 --> 07:16.100] with your physiological [07:16.100 --> 07:19.680] reactions, often involuntary. [07:20.280 --> 07:20.860] This is [07:20.860 --> 07:22.240] not against the law. [07:22.620 --> 07:24.560] This is not covered by normal privacy [07:24.560 --> 07:26.520] law, because privacy law covers your [07:26.520 --> 07:28.880] personal identifying information. [07:29.420 --> 07:30.600] And when you and I entered into [07:30.600 --> 07:32.640] the game, I had a verifiable billing [07:32.640 --> 07:34.680] address. I had a user account [07:34.680 --> 07:36.620] with my authentic identity. [07:36.620 --> 07:38.560] The companies know who we are. That's not [07:38.560 --> 07:40.480] what's at issue. It's the [07:40.480 --> 07:42.880] inferences about our behavior [07:42.880 --> 07:44.220] and our physiology [07:44.220 --> 07:46.520] and our preferences that are really [07:46.520 --> 07:48.560] at risk. So, [07:48.560 --> 07:50.540] it's your mental privacy, not your [07:50.540 --> 07:52.800] identity. So, privacy law [07:52.800 --> 07:54.600] is not the right regime to deal [07:54.600 --> 07:55.660] with this. [07:57.680 --> 07:58.220] It [07:58.220 --> 08:00.500] actually makes me a little bit upset, because [08:00.500 --> 08:02.480] I'm a lawyer, and I'm like, I don't [08:02.480 --> 08:04.600] like it when new technologies [08:04.600 --> 08:06.100] are a misfit with [08:06.100 --> 08:07.640] existing law. [08:13.140 --> 08:14.720] Because there are different [08:14.720 --> 08:16.600] psychological and neurological [08:16.600 --> 08:18.520] impacts than [08:18.520 --> 08:20.280] socializing with people in video game [08:21.040 --> 08:22.240] worlds, [08:22.240 --> 08:24.540] I'd like to review the three characteristics [08:24.540 --> 08:26.940] that make something immersive. [08:27.060 --> 08:28.540] And I think that when you understand [08:28.540 --> 08:29.720] this, you [08:30.460 --> 08:32.420] really get that this is not [08:32.420 --> 08:34.380] flat screen media anymore. [08:34.860 --> 08:36.720] First is presence. [08:37.260 --> 08:38.520] Presence has been studied [08:38.520 --> 08:40.360] in communications departments for [08:40.360 --> 08:42.420] decades, and formally it's [08:42.420 --> 08:44.710] called the illusion of non-mediation. [08:45.740 --> 08:46.380] When you go [08:46.380 --> 08:48.680] into an XR environment, it feels [08:48.680 --> 08:49.760] real. [08:51.200 --> 08:52.540] The first thing [08:52.540 --> 08:54.500] that we do at the Virtual Human [08:54.500 --> 08:56.440] Interaction Lab when we have somebody visit [08:57.100 --> 08:58.460] is I have them walk [08:58.460 --> 08:59.540] the plank. [09:00.200 --> 09:01.920] I put them in an experience [09:01.920 --> 09:03.640] where they are over a chasm and they are [09:03.640 --> 09:05.640] walking along the plank because nothing [09:05.640 --> 09:07.520] brings the feeling of presence [09:07.520 --> 09:08.500] to you [09:09.280 --> 09:11.760] more acutely than feeling that [09:11.760 --> 09:13.760] vertigo and feeling your [09:13.760 --> 09:15.560] heart rate rise and seeing how your [09:15.560 --> 09:17.540] physiology reacts [09:17.540 --> 09:19.540] to the screens and the [09:19.540 --> 09:21.260] spatial audio and the impact [09:21.260 --> 09:23.600] to show you that this is not a Twitter [09:23.600 --> 09:25.700] feed, this is not Facebook, and therefore [09:25.700 --> 09:27.660] maybe we shouldn't be regulating it in the same [09:27.660 --> 09:28.460] way. [09:32.230 --> 09:34.190] Jeremy Bailenson, who is the founding [09:34.190 --> 09:36.290] director of the lab, says we should think of [09:36.290 --> 09:38.330] VR not as a media experience [09:38.330 --> 09:40.330] but closer to an actual [09:40.330 --> 09:42.370] experience. And the [09:42.370 --> 09:44.170] visceral pit in your stomach [09:44.170 --> 09:46.150] you get when you have [09:46.150 --> 09:48.570] fear of heights, [09:48.570 --> 09:50.270] that's what [09:50.270 --> 09:51.350] that is. [09:52.650 --> 09:54.190] And we do that before talking [09:54.190 --> 09:56.250] about empathy and education [09:57.030 --> 09:57.630] and [09:57.630 --> 10:00.030] all of the [10:00.030 --> 10:01.830] other benefits of VR so that people [10:01.830 --> 10:03.570] understand its power. [10:04.010 --> 10:06.030] Second is immersion, and that's when users [10:06.030 --> 10:07.990] really feel like they're in a virtual [10:07.990 --> 10:09.330] environment. [10:09.870 --> 10:11.970] A virtual violin performance has to [10:11.970 --> 10:13.910] have all of the trappings of watching [10:14.090 --> 10:16.290] a performance in an actual theater. [10:16.470 --> 10:18.290] Different views from different seats, [10:18.290 --> 10:20.390] different movement flows through aisles, [10:20.390 --> 10:22.530] spatially oriented sound. [10:22.670 --> 10:23.370] So it's [10:24.630 --> 10:25.840] the way that [10:25.840 --> 10:27.920] XR experiences use your sensory [10:27.920 --> 10:29.960] input to stimulate the [10:29.960 --> 10:32.190] user through light, sound, [10:32.740 --> 10:33.780] sense, tactile [10:33.780 --> 10:35.520] input. [10:36.000 --> 10:37.820] Finally there's embodiment, and this is [10:37.820 --> 10:39.860] my favorite one because it reminds me [10:39.860 --> 10:41.360] of Looney Tunes. [10:42.240 --> 10:43.840] Embodiment is the feeling [10:43.840 --> 10:46.000] like your avatar or your virtual body [10:46.000 --> 10:48.150] is your actual body. [10:48.760 --> 10:49.880] And a researcher [10:49.880 --> 10:51.040] named Mel Slater [10:51.840 --> 10:53.740] took a famous social [10:53.740 --> 10:55.440] psychology experiment called the rubber [10:55.440 --> 10:57.720] hand experiment. And he had [10:57.720 --> 10:59.240] people look down in VR [10:59.240 --> 11:01.700] and see a rubber hand placed in front [11:01.700 --> 11:03.700] of them, and then proceeded to go [11:03.700 --> 11:06.200] Looney Tunes on it, to hit it with a [11:06.200 --> 11:08.500] mallet, to stab it, to really [11:08.500 --> 11:10.240] abuse the virtual arm. [11:10.240 --> 11:12.020] And when you watch the recordings of these, [11:12.020 --> 11:14.920] you can find them on YouTube, people [11:14.920 --> 11:17.020] scream, people jump, [11:17.020 --> 11:18.920] people react like it's their [11:18.920 --> 11:19.440] actual arm. [11:21.500 --> 11:22.440] This is the [11:22.440 --> 11:24.400] power of embodiment, and it can [11:24.400 --> 11:26.740] be beneficial in medical applications [11:26.740 --> 11:28.460] if somebody has phantom limb pain [11:28.460 --> 11:30.500] where you want to show them an arm that is [11:30.500 --> 11:32.460] not theirs, and have their brain actually [11:32.460 --> 11:34.040] associate what happens to that [11:34.040 --> 11:36.020] with their physiognomy. [11:36.120 --> 11:38.520] But it also can place users at risk [11:38.520 --> 11:40.600] considering the violence [11:40.600 --> 11:42.400] and other user harms that we're [11:42.400 --> 11:44.820] starting to see emerge in XR. [11:46.440 --> 11:49.120] So if this feels so real to our bodies, [11:49.120 --> 11:50.920] what does it do to our minds? [11:52.540 --> 11:53.600] Tom Furness, [11:53.600 --> 11:55.020] who is one of the inventors of VR [11:55.020 --> 11:57.100] headsets, says that when something happens [11:57.100 --> 11:59.020] to you in XR, it's like it's [11:59.020 --> 12:00.840] imprinted on your brain in permanent [12:00.840 --> 12:01.780] ink. [12:02.880 --> 12:04.960] The experiences that you have in a virtual [12:04.960 --> 12:07.260] environment are processed through your hippocampus [12:07.260 --> 12:08.740] in the same way that you create [12:08.740 --> 12:11.120] memories, like you [12:11.120 --> 12:13.240] and I sitting here having a conversation [12:13.240 --> 12:14.100] today. [12:15.840 --> 12:16.920] It's more like [12:16.920 --> 12:18.880] entering somebody's living room [12:18.880 --> 12:20.640] and screaming in their face [12:20.640 --> 12:23.060] than it is reading critical [12:23.060 --> 12:24.740] words on Facebook [12:24.740 --> 12:25.960] or Twitter. [12:27.440 --> 12:28.860] Because of how [12:28.860 --> 12:30.980] XR impacts our bodies [12:30.980 --> 12:32.960] and our minds, it's a mistake [12:32.960 --> 12:34.860] to just cut and paste terms [12:34.860 --> 12:37.100] of service from flat screen social [12:37.100 --> 12:38.980] media and presume that [12:38.980 --> 12:41.000] they're going to do the same thing. And that sounds [12:41.000 --> 12:42.920] really obvious, but that's what a lot [12:42.920 --> 12:45.060] of companies are doing these days. [12:46.360 --> 12:47.200] Second, [12:47.200 --> 12:49.240] I promised we'd talk about different stack layers [12:49.240 --> 12:51.100] and moderation vectors in spatial [12:51.100 --> 12:52.160] computing. [12:53.360 --> 12:55.140] What that means in normal people [12:55.140 --> 12:57.140] speak is that when you do content [12:57.140 --> 12:59.300] moderation on Facebook or Twitter, [12:59.300 --> 13:00.940] you look at two things, conduct [13:00.940 --> 13:02.280] and content. [13:03.080 --> 13:05.040] When you move from a 2D [13:05.040 --> 13:07.140] to a 3D environment, you have [13:07.140 --> 13:09.200] to look at content, conduct [13:09.200 --> 13:11.100] and environment. So the [13:11.100 --> 13:13.240] architecture of your world, [13:13.240 --> 13:14.980] like we talked about with the [13:14.980 --> 13:17.500] penises in Lego Universe, [13:17.500 --> 13:19.700] becomes a moderation factor. [13:19.700 --> 13:21.520] One that you don't have [13:21.520 --> 13:23.440] when you have a text-based [13:23.440 --> 13:25.380] social media environment. [13:27.700 --> 13:28.900] That means that there [13:28.900 --> 13:30.800] are actually different ways to [13:30.800 --> 13:33.000] mess with people in [13:33.000 --> 13:34.940] XR worlds. I like [13:34.940 --> 13:36.560] to think about spam. [13:36.680 --> 13:38.800] And how does spam translate into [13:38.800 --> 13:40.960] XR? Pretend [13:40.960 --> 13:42.880] you and I are tired of my [13:42.880 --> 13:44.420] sports car, so we go to a sports [13:44.420 --> 13:46.820] bar. And in the [13:46.820 --> 13:48.460] sports bar, I can buy Liz [13:48.900 --> 13:50.800] a beer. I don't [13:50.800 --> 13:52.900] want Liz coming to my bar [13:52.900 --> 13:55.200] though, or playing on my pool table, [13:55.200 --> 13:56.900] so I buy her so many beers [13:56.900 --> 13:58.920] that it fills up the virtual pool table [13:59.420 --> 14:00.820] and she can't play the [14:00.820 --> 14:01.540] game. [14:03.260 --> 14:04.980] Impeding a user's gameplay [14:04.980 --> 14:06.420] in an XR environment [14:06.420 --> 14:08.600] is the functional equivalent to [14:08.600 --> 14:10.660] spam in a flat screen computing [14:10.660 --> 14:12.520] environment, but it would never be defined [14:12.520 --> 14:14.760] as such if you didn't think about it [14:14.760 --> 14:16.640] in three dimensions. [14:22.160 --> 14:23.360] I'm going to skip a [14:23.360 --> 14:25.160] little bit ahead because we have a lot [14:25.160 --> 14:29.960] to talk about. [14:29.960 --> 14:31.300] Three, platforms [14:31.300 --> 14:33.200] are having a lot of challenges [14:33.880 --> 14:35.340] trying to create trust [14:35.340 --> 14:37.800] and safety regimes and privacy [14:37.800 --> 14:39.640] regimes because of all of [14:39.640 --> 14:42.480] these complicating factors. [14:42.480 --> 14:42.900] This is a [14:42.900 --> 14:44.660] new form of media with [14:44.660 --> 14:47.040] new impacts on your body and mind, you need [14:47.040 --> 14:48.400] new rules. [14:49.340 --> 14:50.900] A lot of people ask, [14:50.900 --> 14:52.920] why didn't Lego just create a filter [14:52.920 --> 14:54.680] to screen out [14:54.680 --> 14:56.180] all the penises? [14:57.160 --> 14:58.780] And the answer was [14:58.780 --> 15:00.640] that most social media content [15:00.640 --> 15:02.860] is moderated by automated [15:02.860 --> 15:04.840] filters in an ex-ante [15:04.840 --> 15:06.920] fashion, so before it ever hits [15:07.400 --> 15:08.580] a flat feed. [15:08.720 --> 15:10.860] Facebook takes 97% of hate [15:10.860 --> 15:13.120] speech off before you ever see it. [15:13.420 --> 15:14.840] And the way that they do this is [15:14.840 --> 15:16.700] by using classifiers. [15:17.180 --> 15:18.000] So it's [15:18.720 --> 15:20.920] basically a huge AI profanity [15:20.920 --> 15:21.940] filter. [15:22.840 --> 15:24.740] This is not possible [15:24.740 --> 15:27.040] in the XR context yet. [15:28.600 --> 15:30.460] We don't have [15:30.460 --> 15:33.040] scalable, affordable, [15:35.080 --> 15:37.180] reliable ways to create [15:38.320 --> 15:39.000] classifiers [15:39.000 --> 15:40.900] for 3D environments. [15:40.960 --> 15:42.700] We don't have the data sets yet [15:42.700 --> 15:45.140] to use. Computer vision is [15:45.140 --> 15:47.060] the way that I think [15:47.060 --> 15:49.080] we're going to be able to do this, but think about [15:49.080 --> 15:50.900] all the different ways you can [15:50.900 --> 15:52.200] describe a cat. [15:53.080 --> 15:54.320] We don't have [15:54.980 --> 15:56.920] publicly available data sets yet to try [15:56.920 --> 15:58.820] to audit what a cat would [15:58.820 --> 16:01.160] actually be in its many forms [16:01.160 --> 16:03.500] and colors and shapes and sizes. [16:04.700 --> 16:06.720] Currently when companies do content [16:06.720 --> 16:08.900] moderation, one, they [16:08.900 --> 16:10.640] just don't do it for XR. [16:10.940 --> 16:12.880] Two, they use a [16:13.860 --> 16:15.540] voice to text transcription [16:15.540 --> 16:17.240] and then run it through social [16:17.240 --> 16:19.040] media filters, which [16:19.040 --> 16:21.180] does not catch context [16:21.180 --> 16:23.180] or behavioral or environmental [16:23.180 --> 16:24.840] factors. Or [16:25.800 --> 16:27.000] three, they try to [16:27.800 --> 16:29.440] use realtime behavioral [16:29.440 --> 16:31.060] tracking, which is [16:31.060 --> 16:33.220] just not scalable and puts the burden [16:33.220 --> 16:34.240] on users. [16:37.080 --> 16:39.200] When you report a violation [16:39.200 --> 16:41.360] in a platform like [16:41.360 --> 16:43.340] Horizon Worlds, it takes one to [16:43.340 --> 16:44.980] two minutes of content [16:44.980 --> 16:47.140] and gives it to the company. [16:47.500 --> 16:49.200] That normally is not [16:49.200 --> 16:51.560] enough to determine the context [16:51.560 --> 16:53.500] of what's actually happening. [16:53.560 --> 16:55.300] And if you watch videos [16:55.300 --> 16:56.880] of people who are reporting [16:56.880 --> 16:58.740] abuse happening, [16:59.200 --> 17:01.240] you often can't tell who is saying what [17:01.240 --> 17:02.960] to whom and what's happening. [17:02.960 --> 17:04.940] Especially in 60 seconds. [17:07.080 --> 17:08.020] Reporting is done [17:08.020 --> 17:09.980] differently as well on every different [17:09.980 --> 17:11.960] platform. So that's like [17:11.960 --> 17:14.060] if we went to [17:14.060 --> 17:16.060] Vegas and you call 911, [17:16.060 --> 17:17.460] but then if you went to Henderson [17:18.180 --> 17:20.440] and you had to send them a fax, [17:20.440 --> 17:21.680] and then if you went to Reno [17:21.680 --> 17:23.680] and you had to show up in person. [17:24.240 --> 17:25.980] When you use public safety [17:25.980 --> 17:27.880] services, if you had to call for help [17:27.880 --> 17:29.880] in a different way in every [17:29.880 --> 17:31.720] different city you went to, you [17:31.720 --> 17:33.760] would never use 911. [17:33.760 --> 17:35.420] And I think [17:35.420 --> 17:37.220] companies are finding that it's the same way [17:37.220 --> 17:39.460] if you don't have standardized safety protocols [17:39.460 --> 17:41.400] across different XR platforms [17:41.400 --> 17:42.740] and XR worlds. [17:46.410 --> 17:48.010] Generative AI looks promising [17:48.010 --> 17:49.850] as a tool to help, but it does not yet [17:49.850 --> 17:51.490] have the discretion or the contextual [17:52.070 --> 17:53.610] understanding necessary to help [17:53.610 --> 17:55.430] with the nuance that will be required [17:56.070 --> 17:57.530] for helping to modify [17:57.530 --> 17:59.430] or govern user behavior. [18:00.110 --> 18:01.530] Realistically, AI based [18:01.530 --> 18:03.490] solutions are several years away from having [18:03.490 --> 18:05.510] sufficient sophistication, enough data [18:05.510 --> 18:07.410] and adequate training to work in spatial [18:07.410 --> 18:08.630] environments. [18:11.700 --> 18:13.920] So, overall, [18:13.920 --> 18:15.140] the differences between [18:15.140 --> 18:17.140] social media and XR [18:17.140 --> 18:18.860] are stark. [18:19.460 --> 18:21.300] And they emerge from [18:21.300 --> 18:23.580] the interactive features of immersive experiences [18:23.580 --> 18:25.200] and the neurological impacts [18:25.200 --> 18:26.880] that this has on users. [18:27.080 --> 18:29.220] And this has resulted in violations of what I [18:29.220 --> 18:31.380] called the right to mental privacy. [18:31.620 --> 18:33.420] Other scholars in Chile have called this [18:33.420 --> 18:35.180] neuro rights. [18:35.300 --> 18:37.180] Brain-computer interface scholars [18:37.180 --> 18:39.420] are talking about a [18:41.460 --> 18:41.780] battle [18:41.780 --> 18:44.200] for your brain. [18:44.560 --> 18:47.180] There's a new book that came out [18:47.180 --> 18:49.040] discussing that. [18:49.040 --> 18:50.440] And they're all [18:50.440 --> 18:52.540] part and parcel of the same thing [18:52.540 --> 18:55.760] where the sanctity of the content of [18:55.760 --> 18:57.240] your thoughts and [18:57.240 --> 18:59.100] the inferences that can be [18:59.100 --> 19:01.380] made about your health [18:59.420 --> 19:01.420] and your sexual [19:01.380 --> 19:03.540] preferences, your political [19:03.540 --> 19:05.680] preferences, your gender, your age, [19:05.680 --> 19:07.400] all of these things we can determine from [19:07.400 --> 19:08.100] data flows [19:09.680 --> 19:11.460] really shouldn't be used [19:11.460 --> 19:13.840] to send you advertisements. [19:14.980 --> 19:15.500] If so, [19:15.500 --> 19:17.360] I think it's actually a failure of the [19:17.360 --> 19:18.400] imagination. [19:25.340 --> 19:27.080] I think that's a good [19:27.080 --> 19:28.380] way to turn it to us. [19:29.940 --> 19:30.780] And one [19:30.780 --> 19:32.620] of the things that really [19:33.440 --> 19:34.620] strikes my mind is [19:34.620 --> 19:36.380] it's not just [19:36.380 --> 19:38.420] the gaming experience. [19:38.420 --> 19:40.680] This is moving beyond [19:40.680 --> 19:42.980] so many different aspects [19:42.980 --> 19:44.600] and potential. [19:44.600 --> 19:46.280] So as Britton was talking about, [19:46.280 --> 19:48.320] moving into the healthcare environment, [19:48.320 --> 19:49.720] that your healthcare [19:49.720 --> 19:51.920] will be delivered potentially [19:52.430 --> 19:54.380] through this data [19:54.380 --> 19:55.620] that is being collected [19:56.410 --> 19:59.000] from your physical reactions. [19:59.000 --> 20:00.540] So the questions [20:00.540 --> 20:02.500] start to become, is this data [20:02.500 --> 20:04.500] painting a correct picture? [20:04.500 --> 20:06.300] Is this the full picture? Is it [20:06.300 --> 20:09.140] picking up on all the nuances? [20:09.440 --> 20:10.480] And who's getting access [20:10.480 --> 20:12.580] to that? And there's [20:12.760 --> 20:14.220] a big difference between [20:14.220 --> 20:16.700] what I may post on [20:16.700 --> 20:18.680] social media at 2am [20:18.680 --> 20:20.500] when I am fired up about [20:20.500 --> 20:22.580] something and what I would actually [20:22.580 --> 20:24.520] say or do. [20:24.520 --> 20:26.180] Or you think of a lot of the, [20:26.180 --> 20:27.980] as we saw during [20:29.140 --> 20:30.420] remote environments [20:30.420 --> 20:32.400] where people were having job [20:32.400 --> 20:34.540] interviews over Zoom [20:34.540 --> 20:35.500] or other [20:37.040 --> 20:38.740] conference calling. [20:38.800 --> 20:40.420] And a lot of it dealt [20:40.420 --> 20:42.240] with, there were companies, and [20:42.240 --> 20:44.540] some states have started taking a harder [20:44.540 --> 20:46.320] look at this, where they were creating [20:46.320 --> 20:48.360] biases. That if you didn't make [20:48.360 --> 20:50.300] the proper amount of eye contact, [20:50.300 --> 20:52.360] or if it looked like you were distracted [20:52.360 --> 20:54.200] during, because [20:54.200 --> 20:55.960] the sensors on the screen [20:55.960 --> 20:58.620] weren't picking this up properly, [20:58.620 --> 21:00.400] that suddenly you were [21:00.400 --> 21:02.160] getting penalized [21:02.160 --> 21:04.260] during your job interview [21:04.260 --> 21:06.200] because you weren't meeting [21:06.200 --> 21:08.460] this set standard. [21:08.460 --> 21:10.640] And so now what happens when we take [21:10.640 --> 21:12.660] that and all of those [21:12.660 --> 21:13.340] AI [21:14.520 --> 21:16.040] data sets that are [21:16.420 --> 21:18.420] expecting certain behaviors, they're not [21:18.420 --> 21:19.360] expecting children [21:20.240 --> 21:22.480] to start drawing inappropriate [21:22.480 --> 21:24.280] things, which [21:24.280 --> 21:27.040] anyone who's been in an elementary school, [21:27.040 --> 21:29.060] you should have expected it. [21:29.060 --> 21:31.520] And you should have expected fart jokes, [21:31.520 --> 21:33.040] you should have expected [21:33.040 --> 21:34.980] lots of poop, [21:34.980 --> 21:37.840] and all those different things. [21:37.840 --> 21:39.300] But what happens [21:39.300 --> 21:41.400] when we're trying to create [21:42.160 --> 21:43.260] protections, [21:43.260 --> 21:44.800] and we're trying to do all this [21:44.800 --> 21:47.060] with a bias towards, like, we're already [21:47.060 --> 21:49.040] starting so far behind. [21:49.040 --> 21:50.960] And now we're taking [21:50.960 --> 21:53.160] all these other little cues of [21:53.160 --> 21:55.360] how does the VR [21:55.860 --> 21:57.720] headset, like how does that experience [21:57.720 --> 21:59.860] know that it wasn't because a [22:00.440 --> 22:01.580] spider crawled across [22:01.580 --> 22:03.580] your foot or someone else in the [22:03.580 --> 22:05.360] room thought it would be hilarious [22:06.040 --> 22:08.440] to tap you on the shoulder [22:08.440 --> 22:10.080] as you were playing the game [22:10.080 --> 22:11.620] or, or, [22:11.620 --> 22:14.100] I'm not saying this has ever happened, I've [22:14.100 --> 22:16.500] definitely never done this to someone [22:16.500 --> 22:18.060] playing, but what happens if you [22:18.060 --> 22:19.700] move furniture while they're [22:19.700 --> 22:21.540] in the middle of and suddenly they're [22:21.540 --> 22:22.960] tripping over something because it was really [22:22.960 --> 22:24.820] freaking funny and [22:24.820 --> 22:26.640] it serves them right [22:26.640 --> 22:28.800] but suddenly that reaction [22:28.800 --> 22:30.920] is getting captured [22:30.920 --> 22:32.940] and how do you know that [22:32.940 --> 22:35.680] I haven't logged in to Britton's account? [22:36.000 --> 22:36.840] That I'm not [22:36.840 --> 22:39.000] playing as her and should I [22:39.000 --> 22:40.760] be so lucky to [22:40.760 --> 22:42.960] have all the sensors pick up [22:42.960 --> 22:44.980] and be like, hmm, you have the same [22:44.980 --> 22:47.000] facial structure, you have all this [22:47.000 --> 22:48.740] and I'm like, yes I do. [22:48.820 --> 22:50.980] Yes, that's right. [22:50.980 --> 22:53.020] Or, you're just off today [22:53.020 --> 22:54.800] or heavens forbid, I think of [22:54.800 --> 22:56.820] all the times I've tried to log in [22:57.500 --> 22:58.940] to my iPhone [22:58.940 --> 23:00.840] and it's like facial recognition, not [23:00.840 --> 23:02.700] recognition, I'm like, wait, what? [23:03.540 --> 23:04.960] No, oh my god, [23:04.960 --> 23:06.980] are you telling me I didn't get enough sleep last night [23:06.980 --> 23:08.780] and suddenly, or [23:08.780 --> 23:11.500] oh, god, is my resting bitch face back? [23:11.500 --> 23:13.160] Like, oh, I didn't mean to do that. [23:13.160 --> 23:15.020] So how do regulators and how [23:15.020 --> 23:17.120] do you work all those different [23:17.120 --> 23:19.000] aspects in? Because we're already [23:19.000 --> 23:20.340] playing catch up. [23:26.240 --> 23:28.000] One of the things that [23:28.000 --> 23:29.620] I like to do when I [23:29.620 --> 23:31.500] talk to regulators is [23:31.500 --> 23:34.060] tell them the type [23:34.060 --> 23:35.560] and quality of information [23:35.560 --> 23:37.860] you can get about a person from these [23:37.860 --> 23:39.660] data flows. And I realize I didn't [23:39.660 --> 23:41.500] mention that before. [23:41.560 --> 23:43.760] There was a study that came out last week [23:43.760 --> 23:46.220] this is the absolute newest information [23:46.220 --> 23:47.240] from UC Berkeley [23:47.920 --> 23:49.120] and the study [23:50.280 --> 23:52.020] said that a machine learning [23:52.020 --> 23:53.960] model was able to accurately [23:53.960 --> 23:56.600] predict the height, weight, age, [23:56.600 --> 23:57.480] marital status [23:58.100 --> 24:01.840] of a user in [24:01.840 --> 24:04.220] XR by the way they moved. [24:07.770 --> 24:08.010] So [24:08.010 --> 24:10.190] I know earlier I said that it doesn't [24:10.190 --> 24:12.090] go into the realm of personal identifying [24:12.090 --> 24:14.070] information, but initial studies [24:14.070 --> 24:16.130] are starting to show that [24:16.130 --> 24:17.370] you can actually tell somebody's [24:18.150 --> 24:20.170] foot size by the way [24:20.170 --> 24:21.130] they move. [24:21.130 --> 24:23.050] The lab at [24:23.050 --> 24:24.370] Berkeley and the lab at Stanford [24:24.850 --> 24:26.510] have had some [24:27.130 --> 24:28.890] similar studies and Berkeley [24:28.890 --> 24:30.830] came out with a study that [24:30.830 --> 24:31.170] said [24:31.170 --> 24:34.750] there was 100 seconds [24:34.750 --> 24:36.890] of recorded data you could [24:36.890 --> 24:38.870] uniquely identify a person [24:38.870 --> 24:40.830] in XR with 94% [24:40.830 --> 24:41.530] accuracy [24:41.530 --> 24:44.650] from a data set of 50,000 [24:44.650 --> 24:45.810] people. [24:46.950 --> 24:48.610] Stanford did this earlier [24:48.610 --> 24:50.730] but the data set was [24:51.170 --> 24:52.710] a couple thousand people, it wasn't [24:52.710 --> 24:54.830] statistically significant. [24:55.350 --> 24:57.590] 50,000 people. [24:57.590 --> 24:58.570] That might be [24:58.570 --> 25:00.870] you in the Taylor Swift crowd [25:00.870 --> 25:02.610] by the way that you tilt [25:02.610 --> 25:04.730] your head and point can be uniquely [25:04.730 --> 25:07.390] identified in an immersive environment. [25:08.310 --> 25:09.030] The type [25:09.030 --> 25:10.830] of medical [25:10.830 --> 25:12.710] and sort of human rights [25:12.710 --> 25:14.850] oriented information that I'm concerned about [25:14.850 --> 25:16.850] really is [25:16.850 --> 25:18.350] tied into eye tracking. [25:19.270 --> 25:20.890] So the type of information you can get [25:20.890 --> 25:22.710] by somebody's pupil reaction [25:23.550 --> 25:24.770] you can tell who [25:24.770 --> 25:26.050] somebody is sexually attracted to [25:27.010 --> 25:28.830] you can tell whether or not somebody is likely [25:28.830 --> 25:29.690] to be telling the truth [25:30.430 --> 25:32.650] and you can also tell whether or not somebody [25:32.650 --> 25:35.110] shows preclinical signs of certain [25:35.110 --> 25:36.390] medical ailments [25:36.390 --> 25:38.650] like schizophrenia, ADHD [25:38.650 --> 25:40.910] Parkinson's and Huntington's. [25:41.010 --> 25:42.370] These are preclinical signs [25:42.370 --> 25:44.930] these are things you may not know about yourself [25:44.930 --> 25:46.710] yet, your doctor may not [25:46.710 --> 25:47.810] know about you yet [25:47.810 --> 25:51.130] and a company [25:51.130 --> 25:52.790] if they were to [25:52.790 --> 25:54.870] parse the data set would have [25:54.870 --> 25:56.950] access to that information. [25:57.630 --> 25:59.050] This is what I mean [25:59.050 --> 26:01.290] by the right to mental privacy. [26:02.150 --> 26:02.810] It's not [26:02.810 --> 26:04.670] just the content [26:04.670 --> 26:06.630] of your thoughts, it's the [26:06.630 --> 26:08.830] implications that can be made about [26:08.830 --> 26:10.890] who you uniquely are [26:10.890 --> 26:13.030] your medical conditions [26:13.030 --> 26:15.310] and the type of behaviors that you may [26:15.310 --> 26:17.830] choose to engage in or not. [26:19.950 --> 26:21.130] One of the things [26:21.130 --> 26:23.130] we're seeing is the states are starting [26:23.130 --> 26:24.550] to try to play catch up [26:24.550 --> 26:27.310] because, keeping in mind [26:27.310 --> 26:28.790] data brokers [26:28.790 --> 26:30.990] are taking this information and [26:30.990 --> 26:33.350] creating these pictures [26:33.350 --> 26:34.850] and they're [26:34.850 --> 26:36.670] selling them. And [26:36.670 --> 26:39.410] the regulations really are not [26:39.410 --> 26:41.370] keeping up [26:41.370 --> 26:43.230] with who should have access [26:43.230 --> 26:46.090] to this and how do we protect it. [26:46.290 --> 26:47.710] And it's also [26:47.710 --> 26:49.270] what happens if [26:49.270 --> 26:51.410] it gets wrong. [26:51.630 --> 26:53.270] That one [26:53.270 --> 26:55.450] out of 50,000 [26:55.450 --> 26:57.370] it's wrong. [26:57.610 --> 26:59.410] But yet this is now [26:59.410 --> 27:01.410] flagged in a file and it's [27:01.410 --> 27:03.390] you know as a child you're told this will go [27:03.390 --> 27:05.230] down in your permanent record. [27:05.850 --> 27:07.530] Well how do we fix that [27:07.530 --> 27:09.430] permanent record when it's our [27:09.430 --> 27:11.530] biometric and it's [27:11.530 --> 27:13.370] our mental information [27:13.370 --> 27:15.210] that, oh wait, you know, [27:15.210 --> 27:17.650] objection your honor. I was just [27:17.650 --> 27:19.810] having a bad day and that wasn't me. [27:20.450 --> 27:20.970] Or [27:20.970 --> 27:22.890] the, you know, it [27:23.350 --> 27:25.470] that was someone else and that [27:25.470 --> 27:27.510] piece of information. So it's not the [27:27.510 --> 27:29.170] entire picture that gets off. [27:29.170 --> 27:31.530] It's just one piece. How do you fix [27:31.530 --> 27:32.350] that? [27:36.970 --> 27:38.810] I don't have a good answer for that. [27:38.810 --> 27:40.430] I wish I did. [27:40.430 --> 27:42.190] Because right now [27:42.190 --> 27:44.250] these type of data sets aren't [27:44.250 --> 27:45.970] auditable by the public. [27:47.070 --> 27:48.410] In certain jurisdictions [27:48.410 --> 27:50.330] people are proposing a right to have [27:50.330 --> 27:52.430] all your user information about you [27:52.430 --> 27:54.510] available, but these are [27:55.610 --> 27:56.590] profiles made [27:56.590 --> 27:58.090] to sell you something. [27:58.390 --> 28:00.110] It's not an essential part of your [28:00.110 --> 28:02.130] user account and things. [28:02.870 --> 28:04.530] Another thing I tell regulators [28:04.530 --> 28:06.430] is that you [28:06.430 --> 28:08.390] can't outsmart the machine. [28:08.390 --> 28:10.130] Because a lot of them think that [28:11.310 --> 28:11.830] somebody [28:12.490 --> 28:14.650] advertising something to me, that's obvious. [28:14.650 --> 28:16.330] That's not [28:16.330 --> 28:18.630] I'm not going to fall for that. [28:18.890 --> 28:20.290] In XR environments you can actually [28:20.290 --> 28:21.870] switch out elements [28:22.390 --> 28:24.910] in the environment at blink rate. [28:27.550 --> 28:28.070] It's [28:28.070 --> 28:30.590] been done in experimental context [28:30.590 --> 28:32.250] but there's nothing saying that you could [28:32.250 --> 28:34.110] have somebody with a can of soda [28:34.110 --> 28:36.330] and switching out Coke and [28:36.330 --> 28:37.850] Pepsi and [28:37.850 --> 28:40.310] tracking somebody's more [28:40.310 --> 28:43.130] positive reaction to one versus the other. [28:43.510 --> 28:44.350] It's kind of like [28:44.350 --> 28:46.330] that social psychology experiment where [28:46.330 --> 28:48.490] you're told to count the number of [28:48.490 --> 28:50.390] basketball passes and so you [28:50.390 --> 28:52.830] miss the gorilla walking in the background. [28:53.930 --> 28:54.390] It's [28:54.550 --> 28:56.070] a similar type of a dynamic [28:56.070 --> 28:58.470] where people think they're smarter [28:58.470 --> 29:00.370] than the system, but the system [29:00.370 --> 29:02.250] works faster than you can [29:02.250 --> 29:04.350] articulate your preferences or [29:04.350 --> 29:06.450] requests cognitively. [29:07.670 --> 29:08.010] And that's [29:08.890 --> 29:11.010] a really kind of funky, scary thing. [29:11.110 --> 29:12.330] If you watch [29:13.190 --> 29:13.870] sports [29:13.870 --> 29:16.430] and Liz is here and I'm back in [29:16.430 --> 29:18.410] California, Liz [29:18.410 --> 29:20.590] and I will actually see different ads [29:20.590 --> 29:22.110] behind the players. [29:22.950 --> 29:24.470] Augmented advertising is [29:24.470 --> 29:26.350] already a thing and [29:26.350 --> 29:27.610] it's based on your [29:28.530 --> 29:30.190] zip code. [29:31.050 --> 29:32.470] And somebody in the stadium [29:32.470 --> 29:34.410] will see even different [29:34.410 --> 29:36.450] ads than Liz or I would see [29:36.450 --> 29:38.230] on different network channels. [29:38.250 --> 29:40.190] They sell this space. [29:40.410 --> 29:42.530] Advertisement in an immersive [29:42.530 --> 29:43.950] context is more about [29:45.090 --> 29:46.970] targeting and personalization [29:46.970 --> 29:48.450] than it is [29:48.450 --> 29:50.790] in an offline context. And that makes sense, [29:50.790 --> 29:51.890] but the capacity [29:52.470 --> 29:54.590] and the ability to really [29:54.590 --> 29:55.790] target it to people [29:56.550 --> 29:58.610] makes XR the most persuasive [29:58.610 --> 30:00.510] medium that we've [30:00.510 --> 30:01.710] developed so far. [30:03.010 --> 30:04.490] And one of the things that [30:04.490 --> 30:06.190] we saw and we keep going back to is [30:07.170 --> 30:08.690] the data sets. [30:08.690 --> 30:10.390] What is it getting based on? Where [30:10.390 --> 30:12.370] are those data sets getting built from [30:12.370 --> 30:14.330] that are creating some of these [30:14.330 --> 30:16.010] models and [30:16.710 --> 30:18.250] being able to audit it [30:18.250 --> 30:20.610] and being able to audit and get to your [30:20.610 --> 30:22.430] information and make sure, like, hey, [30:22.430 --> 30:25.010] is this the picture that I see? [30:25.210 --> 30:26.390] And basing it [30:26.390 --> 30:28.530] on a bias towards, [30:28.530 --> 30:30.430] well, English [30:30.430 --> 30:32.790] speaking, US [30:32.790 --> 30:34.510] based things, when you're [30:34.510 --> 30:36.670] talking about an immersive experience [30:36.670 --> 30:38.450] that is intended to [30:38.450 --> 30:40.610] open up the world. [30:40.610 --> 30:42.470] Like, the idea is that you [30:42.470 --> 30:44.430] were playing and you were experiencing [30:44.430 --> 30:46.530] things with people [30:46.530 --> 30:48.730] in any different physical [30:48.730 --> 30:50.610] location. So now [30:50.610 --> 30:52.370] how do you build the [30:52.370 --> 30:55.010] rules and even train [30:55.010 --> 30:56.270] that it's getting it right [30:56.270 --> 30:58.390] when not [30:58.390 --> 31:00.410] everyone's going to look the same, react the same [31:00.410 --> 31:02.530] and different cultures have built [31:02.530 --> 31:03.790] in different [31:05.370 --> 31:06.490] mannerisms. [31:06.890 --> 31:08.590] And as we saw with a lot [31:08.590 --> 31:10.390] of the, you know, again, [31:10.390 --> 31:12.390] even just trying to unlock your iPhone, [31:13.050 --> 31:14.570] heavens forbid, you're [31:14.950 --> 31:16.490] a woman of color. [31:16.490 --> 31:18.390] Because guess what? All of [31:18.390 --> 31:20.330] the modeling that was based [31:20.330 --> 31:22.210] on was not you. [31:22.350 --> 31:24.450] Not even a little bit. And so they had [31:24.450 --> 31:26.850] to go back and say, oops, our bad. [31:26.910 --> 31:28.310] Well now, what do we do when we're [31:28.310 --> 31:30.470] talking about collecting all this data [31:30.470 --> 31:31.930] and it's like, on [31:32.310 --> 31:34.270] the reactions of [31:34.270 --> 31:36.410] you know, we talk about a lot in [31:36.410 --> 31:38.470] the medical world, a lot [31:38.470 --> 31:39.430] of the early [31:40.330 --> 31:42.470] training was done and it wasn't [31:42.470 --> 31:44.530] done based on women's [31:44.530 --> 31:46.490] bodies. And it wasn't based [31:46.490 --> 31:49.070] on... so it's like, oh, well, [31:49.070 --> 31:50.350] you're going to react to this [31:50.350 --> 31:52.450] medication in the following manners, or these [31:52.450 --> 31:54.390] are the following symptoms. It's like, but [31:55.930 --> 31:57.290] no. [31:57.550 --> 31:58.710] I am not. [31:58.710 --> 32:00.410] So now you've got the VR and [32:00.410 --> 32:02.530] it's assuming, heavens forbid, [32:02.530 --> 32:04.510] I don't want to be [32:05.990 --> 32:06.670] a redhead [32:06.670 --> 32:08.330] middle aged woman in [32:08.330 --> 32:10.530] my, like, I want to [32:10.530 --> 32:12.710] be a kick ass [32:12.710 --> 32:14.410] like ninja [32:14.410 --> 32:16.430] dragon, whatever the hell [32:16.430 --> 32:18.390] I want. So [32:18.390 --> 32:20.330] how are we going to start [32:20.330 --> 32:22.370] building that and work on regulators when you've got [32:22.370 --> 32:24.490] because, again, [32:24.490 --> 32:26.670] heavens forbid, heavens forbid [32:26.670 --> 32:27.770] that [32:28.450 --> 32:30.390] Australia, the US, New Zealand [32:30.390 --> 32:31.310] like, insert [32:32.150 --> 32:34.290] country, get the upper [32:34.290 --> 32:35.790] hand in being [32:36.330 --> 32:38.170] the regulator who [32:38.170 --> 32:40.330] sets the rules for this. [32:40.330 --> 32:42.370] Because, oh no, we can't let [32:42.370 --> 32:44.290] Canada do that. I mean, they're just nice [32:44.290 --> 32:45.750] and maple syrup. [32:46.450 --> 32:48.510] So how do you broach [32:48.510 --> 32:49.670] those conversations [32:50.610 --> 32:52.570] when you're going all over the world [32:52.570 --> 32:54.410] and talking to them and saying, no, no, no, [32:54.410 --> 32:56.350] we need to let this country take the lead [32:56.350 --> 32:57.990] or play nice with them [32:58.410 --> 32:59.950] and not share the data. [33:06.860 --> 33:08.200] If y'all notice, I give [33:08.200 --> 33:10.460] Britton all the really hard questions [33:10.460 --> 33:12.180] because this is what she gets for being [33:12.180 --> 33:14.120] so smart and being a professor, [33:14.120 --> 33:15.720] shaping young minds [33:15.720 --> 33:18.200] that are creating this future. So, sorry. [33:18.220 --> 33:20.120] But those aren't young minds, those are like [33:20.120 --> 33:22.020] crusty old minds. That's why I'm like, [33:22.020 --> 33:23.640] hmm, about this. [33:25.080 --> 33:26.100] A lot of people [33:26.100 --> 33:28.240] in Europe argue that GDPR [33:28.240 --> 33:29.800] should cover it. [33:29.960 --> 33:32.000] They argue that Article 1 of GDPR [33:32.000 --> 33:34.060] should cover all of the privacy protections [33:34.060 --> 33:36.180] for XR and then we're done. [33:36.460 --> 33:38.220] I take issue with that. [33:38.780 --> 33:40.260] Because these are new types [33:40.260 --> 33:41.800] of data flows and [33:41.800 --> 33:44.060] they're not really anticipated by [33:44.060 --> 33:46.000] the regulatory structure. [33:47.220 --> 33:48.340] And they're [33:48.340 --> 33:50.020] more akin to medical data [33:50.020 --> 33:51.640] to me than they are personal [33:51.640 --> 33:53.540] identifying information. [33:56.800 --> 33:58.340] The Future on Privacy [33:58.340 --> 34:00.520] Forum just did a [34:01.700 --> 34:02.980] really optimistic [34:02.980 --> 34:04.220] study where [34:04.220 --> 34:06.480] they determined that biometric laws [34:06.480 --> 34:08.260] are, biometric laws in the US [34:08.260 --> 34:10.960] are covered by state based law. [34:10.960 --> 34:12.520] And if you have a business [34:12.520 --> 34:14.020] that creates XR [34:14.720 --> 34:16.240] products, you know this because there are [34:16.240 --> 34:18.020] certain states you're really, really worried [34:18.020 --> 34:20.020] about selling in and certain states [34:20.020 --> 34:22.140] you may choose to not release your product [34:22.140 --> 34:24.320] in. And those may be [34:24.320 --> 34:26.100] Illinois, Washington, or [34:26.100 --> 34:26.940] Texas. [34:28.200 --> 34:29.880] There are other states that have privacy [34:29.880 --> 34:32.180] laws, not biometric laws, but [34:33.720 --> 34:34.240] you [34:34.240 --> 34:36.320] think about, sometimes I talk to regulators [34:36.320 --> 34:38.180] and I ask them if they've tried [34:38.180 --> 34:39.480] Smell-O-Vision. [34:40.300 --> 34:42.260] And they look at me like I'm nuts. [34:42.580 --> 34:44.200] And I tell them that in the lab [34:44.200 --> 34:46.120] and in commercial context now, [34:46.120 --> 34:48.600] you can actually go before a [34:48.600 --> 34:50.640] campfire in XR and you don't [34:50.640 --> 34:52.560] just hear the campfire and you don't just [34:52.560 --> 34:55.200] see the campfire, you can actually [34:55.200 --> 34:57.060] smell the charcoal [34:57.060 --> 34:58.900] and in the lab we put a little [34:58.900 --> 35:01.020] toasted marshmallow in there as well just for [35:01.780 --> 35:02.500] fun. [35:02.500 --> 35:07.860] You can go into a garden and smell flowers. [35:07.860 --> 35:09.240] These type of inputs [35:09.240 --> 35:10.920] are really not anticipated [35:10.920 --> 35:14.080] by law. [35:12.120 --> 35:14.120] Also, [35:14.080 --> 35:17.280] there was a [35:17.280 --> 35:18.080] company that [35:19.900 --> 35:21.020] recently released a [35:21.900 --> 35:22.460] breath [35:22.460 --> 35:25.120] gauge for a meditation app. [35:25.120 --> 35:27.320] And my question for them is [35:27.320 --> 35:29.140] is that biometric data? [35:29.140 --> 35:31.120] It could be body-based data [35:31.120 --> 35:32.340] based on some of the states [35:32.880 --> 35:34.580] so it would be regulated as such. [35:34.580 --> 35:36.880] In other states, it wouldn't be. [35:36.880 --> 35:38.980] I think the way that I make regulators [35:38.980 --> 35:40.440] understand this is I [35:40.440 --> 35:42.120] try to actually bring them into the [35:42.120 --> 35:43.460] developer's mindset [35:43.940 --> 35:45.400] and bring them into [35:47.680 --> 35:50.500] the chair of somebody who has to make a [35:50.500 --> 35:52.540] decision about [35:52.540 --> 35:54.620] selling and producing these products [35:54.620 --> 35:56.560] and doing so in an [35:56.560 --> 36:03.830] environment that is ambiguous at best. [36:03.830 --> 36:07.070] Some countries are better about it than others. [36:07.070 --> 36:07.670] Australia [36:07.670 --> 36:09.450] has an e-safety commissioner's office [36:09.450 --> 36:11.750] so they're very, very motivated. [36:11.750 --> 36:13.350] New Zealand is very concerned about public [36:13.350 --> 36:15.310] safety in the online [36:15.310 --> 36:17.370] context. France has had [36:17.790 --> 36:19.530] a national metaverse strategy. [36:19.950 --> 36:21.110] China and [36:21.110 --> 36:22.630] Colombia and [36:23.990 --> 36:25.270] one other country [36:25.270 --> 36:27.210] have started to hold trials in the [36:27.210 --> 36:29.350] metaverse. Don't do that [36:29.350 --> 36:31.730] by the way. That's not a good idea. [36:32.990 --> 36:35.350] Countries are starting to engage in little [36:35.350 --> 36:36.930] ways. I think [36:36.930 --> 36:38.330] Bahamas had [36:40.030 --> 36:41.170] an embassy [36:41.170 --> 36:43.510] in the metaverse. [36:44.830 --> 36:45.610] I think [36:45.610 --> 36:47.370] the way to actually get to these countries is [36:47.370 --> 36:49.490] to talk to [36:49.490 --> 36:51.550] them more about their economies [36:51.550 --> 36:53.310] and corporate environments and that [36:53.310 --> 36:55.570] should get them more motivated rather [36:55.570 --> 36:57.390] than talking about digital diplomacy [36:57.390 --> 36:58.510] and [36:59.630 --> 37:01.570] kind of a new [37:01.570 --> 37:03.130] sovereign space or other [37:03.130 --> 37:05.490] type of nonsense things. [37:06.330 --> 37:07.450] When you think [37:07.450 --> 37:09.730] about how even in the U.S. [37:09.930 --> 37:10.530] a lot of our [37:11.790 --> 37:13.970] laws that touch on [37:13.970 --> 37:14.910] things that we deal with [37:14.910 --> 37:18.050] day to day, the DMCA, [37:18.050 --> 37:19.850] CFA, a lot of them came [37:19.850 --> 37:21.930] because members of Congress [37:21.930 --> 37:23.830] for example were afraid that their [37:23.830 --> 37:25.890] blockbuster rental history [37:25.890 --> 37:28.470] was going to be published. [37:28.990 --> 37:29.950] Oh yeah. [37:29.950 --> 37:31.670] Or they saw a movie [37:31.670 --> 37:34.470] and they said, you know what? [37:34.470 --> 37:36.490] War games. Huh. [37:36.530 --> 37:37.510] Huh. [37:37.510 --> 37:39.470] So I love [37:39.470 --> 37:41.310] that idea of putting it in that [37:41.310 --> 37:43.330] personal context because [37:43.330 --> 37:45.670] unfortunately that does [37:45.670 --> 37:46.950] ground it for them [37:47.370 --> 37:49.310] and show them because so often we [37:49.310 --> 37:51.350] spend time chasing the technology [37:51.350 --> 37:53.630] like oh we have to react [37:53.630 --> 37:55.150] to this when it's like no [37:55.150 --> 37:57.290] let's look at what it's [37:57.290 --> 37:59.610] actually doing. What is it doing at its core? [37:59.610 --> 38:01.750] What is the actual concern? [38:01.770 --> 38:03.750] And then how do we build from there? [38:03.750 --> 38:05.510] So I love that you're [38:05.510 --> 38:07.510] getting them that hands on but how do y'all [38:07.510 --> 38:09.590] scale that kind [38:09.590 --> 38:11.350] of advocacy or how do we [38:11.350 --> 38:13.450] as a community [38:13.450 --> 38:15.630] of users who are concerned [38:15.630 --> 38:17.170] about or builders [38:17.170 --> 38:19.770] how do we take that [38:19.770 --> 38:21.790] idea and that approach [38:21.790 --> 38:23.070] and scale it [38:23.070 --> 38:24.070] beyond [38:25.270 --> 38:27.110] and bring it down to that [38:27.110 --> 38:29.470] local level or that regional level [38:29.470 --> 38:30.750] of hey [38:31.370 --> 38:33.470] do this, see this, experience this and know [38:33.470 --> 38:35.670] that this is going to impact what you're doing. [38:36.490 --> 38:37.550] I think you don't [38:37.550 --> 38:39.610] use law. I want [38:39.610 --> 38:41.210] to call out someone in the front row [38:41.210 --> 38:43.450] Caroline Cinders who's [38:43.450 --> 38:45.690] research about how to use effective [38:45.690 --> 38:47.570] design to create healthy online [38:47.570 --> 38:49.510] communities and environments is [38:49.510 --> 38:50.810] awesome. [38:52.170 --> 38:53.270] Yay! [38:53.910 --> 38:55.290] So looking at [38:55.290 --> 38:57.410] design based principles, [38:57.410 --> 38:59.290] looking at social psychology [38:59.290 --> 39:01.590] about how we create cultures [39:02.570 --> 39:03.590] some of my [39:03.590 --> 39:05.690] very boring academic work focuses [39:05.690 --> 39:07.850] on the theory [39:07.850 --> 39:09.590] of code as law and how [39:09.590 --> 39:11.910] that has ignored norms. [39:12.210 --> 39:13.650] And norms are the way that we create [39:13.650 --> 39:15.830] culture and we self regulate our own [39:15.830 --> 39:17.650] online environments. [39:17.650 --> 39:19.370] And try not to place [39:19.370 --> 39:21.250] the full burden on [39:22.210 --> 39:23.390] users but [39:23.390 --> 39:25.490] distribute it collectively amongst [39:25.770 --> 39:27.150] a whole community. [39:28.170 --> 39:29.630] It seems [39:29.630 --> 39:31.530] like a cop out answer because I'm a lawyer. [39:32.830 --> 39:33.790] But this [39:33.790 --> 39:35.510] isn't a legal problem I think. [39:35.510 --> 39:37.710] It's more of a [39:37.710 --> 39:40.170] social problem and an issue [39:40.170 --> 39:41.810] of science not catching up [39:41.810 --> 39:45.510] to politics. [39:45.510 --> 39:45.910] Well and I [39:45.910 --> 39:49.050] think that also putting it on the users [39:50.090 --> 39:50.930] overlooks [39:50.930 --> 39:52.370] the builders, the makers. [39:52.370 --> 39:54.650] Like if we are intentional [39:54.650 --> 39:55.990] from the design aspect [39:55.990 --> 39:57.770] and when I say we, I mean the people [39:57.770 --> 39:59.250] at the companies who are building it [39:59.570 --> 40:01.770] from product to engineering [40:01.770 --> 40:03.670] to all that understanding and [40:03.670 --> 40:05.410] seeing that big picture [40:05.990 --> 40:07.670] and knowing when to call it [40:07.670 --> 40:09.490] oh this is going to impact this. [40:09.750 --> 40:11.770] Let's be intentional. Let's bring them into the [40:11.770 --> 40:13.750] conversation. So as [40:14.330 --> 40:15.690] a DEF CON community [40:15.690 --> 40:17.950] I think that's one of the easiest ways [40:17.950 --> 40:19.710] to start. I was like [40:19.710 --> 40:20.990] start the change. [40:21.210 --> 40:23.490] And I think we can also be really [40:23.490 --> 40:25.990] intentional about how we design [40:25.990 --> 40:27.950] these products and who [40:27.950 --> 40:29.990] we are accommodating [40:29.990 --> 40:31.910] when we design them. [40:32.730 --> 40:33.910] How many people [40:33.910 --> 40:36.690] here are familiar with simulator sickness? [40:36.850 --> 40:38.270] Probably a bunch. [40:38.310 --> 40:39.770] If you're not, it's [40:39.770 --> 40:42.090] if you've gone into a VR world and you [40:42.090 --> 40:43.990] feel nauseous and sick to your [40:43.990 --> 40:45.710] stomach and really disoriented [40:45.710 --> 40:48.370] that's simulator sickness. [40:48.890 --> 40:50.310] In the earlier [40:50.310 --> 40:52.050] iterations of headsets [40:52.770 --> 40:54.290] more women than men [40:54.290 --> 40:56.610] reported simulator sickness. [40:56.610 --> 40:58.330] People started to write about this. [40:58.330 --> 41:00.090] Some said it was just my delicate [41:00.090 --> 41:01.190] female constitution [41:01.830 --> 41:04.030] and the fact that I may [41:04.950 --> 41:06.590] produce babies one day [41:06.590 --> 41:08.310] or the fact that I [41:08.310 --> 41:10.290] don't play first person shooter [41:10.290 --> 41:12.510] games. It's all bullshit. [41:12.910 --> 41:14.270] What actually was happening [41:14.270 --> 41:16.030] was that the headset was designed [41:16.030 --> 41:18.310] for an average male form [41:18.310 --> 41:20.150] between 5'10 and [41:20.150 --> 41:22.410] 6 feet tall. There is one [41:22.410 --> 41:24.330] measurement that is essential to preventing [41:24.330 --> 41:26.450] you getting simulator sickness. [41:26.450 --> 41:28.890] It's the interpupillary distance. [41:28.890 --> 41:29.870] The distance between [41:30.410 --> 41:32.370] your eyes. If you go to [41:32.370 --> 41:34.410] the optometrist, they measure this to put you [41:34.410 --> 41:36.470] in a pair of prescription glasses. [41:36.890 --> 41:38.270] The interpupillary [41:38.270 --> 41:40.670] distance on early VR headsets [41:40.670 --> 41:42.270] was not adjustable. [41:42.570 --> 41:44.610] So it was akin to putting 51% [41:44.610 --> 41:46.490] of the population in the wrong [41:46.490 --> 41:48.050] pair of glasses. [41:48.710 --> 41:50.590] No wonder women were reporting [41:50.590 --> 41:52.990] sickness and discomfort. [41:53.130 --> 41:54.610] And there were studies following that [41:54.610 --> 41:56.730] by Jessica Outlaw that actually showed [41:56.730 --> 41:58.630] people that women weren't returning to [41:58.630 --> 42:01.070] VR environments because they felt uncomfortable. [42:01.830 --> 42:02.770] There were similar [42:02.770 --> 42:04.270] studies that came out. [42:04.270 --> 42:06.930] One from the MIT Media Lab [42:06.930 --> 42:08.930] where there was a researcher doing her [42:08.930 --> 42:10.590] master's thesis work in [42:10.590 --> 42:12.790] Nairobi. She took Oculus Go [42:12.790 --> 42:14.390] to try to put it on [42:15.230 --> 42:16.430] Kenyan entrepreneurs [42:16.430 --> 42:18.730] and she found that with the women [42:18.730 --> 42:20.430] 50% of the time [42:20.430 --> 42:22.370] the strap snapped [42:22.370 --> 42:25.210] when she tried to fit it over their hair. [42:25.710 --> 42:27.130] 50% [42:27.130 --> 42:28.570] of the time. [42:28.670 --> 42:30.730] She ended up deviating from [42:30.730 --> 42:32.710] her master's thesis to design an [42:32.710 --> 42:35.190] adjustable fucking strap. [42:36.050 --> 42:36.810] Easy [42:36.810 --> 42:39.010] fix if you think about the wide [42:39.010 --> 42:40.850] beautiful variety of people [42:40.850 --> 42:42.750] who you want to be [42:42.750 --> 42:44.790] your customers. It's smart [42:44.790 --> 42:46.770] to think about the wide diversity [42:46.770 --> 42:49.090] of body shapes and types. [42:49.970 --> 42:51.130] Disabled populations [42:51.130 --> 42:52.890] are some of the earliest adopters [42:52.890 --> 42:54.510] of XR hardware. [42:54.830 --> 42:56.730] And they are the last people that many [42:56.730 --> 42:58.790] companies think about when they design [42:58.790 --> 43:01.190] these products and services. [43:01.530 --> 43:02.770] Oculus didn't change [43:02.770 --> 43:04.790] their vantage point to make [43:04.790 --> 43:06.630] it adjustable whether or not I was [43:06.630 --> 43:08.830] standing or sitting until [43:08.830 --> 43:10.350] update 31. [43:11.130 --> 43:12.670] And it came on the heels of a Scientific [43:12.670 --> 43:14.670] American article that showed [43:14.670 --> 43:16.370] that somebody had to modify an [43:16.370 --> 43:18.750] Xbox controller to use a rock [43:18.750 --> 43:20.670] climbing game. They were [43:20.830 --> 43:22.590] a user with muscular dystrophy and so [43:22.590 --> 43:24.650] they really wanted to know what rock climbing [43:24.650 --> 43:25.830] felt like. [43:26.490 --> 43:28.370] And they had to modify it themselves [43:28.370 --> 43:30.910] because the game wouldn't accommodate them. [43:32.970 --> 43:34.590] This is something that we [43:34.590 --> 43:37.130] can do as a community. [43:38.370 --> 43:38.670] So [43:38.670 --> 43:40.410] when you're doing your [43:40.410 --> 43:43.070] prototype testing or your [43:43.070 --> 43:44.570] platform building or [43:44.570 --> 43:47.110] your experiential [43:47.110 --> 43:49.050] design, think [43:49.050 --> 43:51.090] about all of the people you want [43:51.090 --> 43:53.070] in your world. And then go [43:53.070 --> 43:53.910] bigger. [43:54.930 --> 43:57.370] Yeah. And I think bringing it back [43:57.370 --> 43:59.070] to what [43:59.070 --> 44:01.070] Britton was talking about before is like that full [44:01.070 --> 44:02.950] stack of an [44:02.950 --> 44:04.890] understanding it's not [44:05.670 --> 44:07.130] a level playing field. And by that [44:07.130 --> 44:08.790] I mean it's not a flat [44:08.790 --> 44:10.770] screen. And so [44:10.770 --> 44:12.450] as you're building and as you're [44:12.450 --> 44:14.670] designing, taking all of those different [44:14.670 --> 44:16.830] aspects and [44:16.830 --> 44:18.810] also incorporating it [44:18.810 --> 44:21.170] into the conversations [44:21.170 --> 44:23.130] on the biometric data. [44:23.130 --> 44:24.810] What exactly? So we're not [44:24.810 --> 44:27.090] using those same definitions. [44:27.550 --> 44:29.390] It's not the same game. [44:29.390 --> 44:30.910] And being, again, [44:30.910 --> 44:32.490] intentional in how you're [44:32.490 --> 44:34.890] building, how you're designing, but also how you're [44:34.890 --> 44:36.610] approaching it. And [44:36.610 --> 44:38.830] be it with the conversations you're having [44:38.830 --> 44:40.970] with regulars as you're watching [44:40.970 --> 44:42.770] that. I think really [44:42.770 --> 44:44.910] just kind of bringing it all. This is [44:44.910 --> 44:46.890] not, you know, this is [44:46.890 --> 44:48.850] an immersive environment. It is the [44:48.850 --> 44:51.250] entire all encompassing environment. [44:51.670 --> 44:52.790] And there's some best [44:52.790 --> 44:54.950] practices that you can follow as [44:54.950 --> 44:56.130] developers [44:56.740 --> 44:59.010] or even as security experts. [44:59.090 --> 45:00.310] A best practice [45:00.830 --> 45:02.910] for XR is to actually [45:02.910 --> 45:05.330] keep the data on device. [45:06.230 --> 45:06.630] That [45:06.630 --> 45:08.470] means that you have to get a warrant to [45:08.470 --> 45:09.890] access the data. [45:10.550 --> 45:12.470] The problem with that is [45:12.470 --> 45:14.810] it really impedes [45:14.810 --> 45:16.490] your processing speed. It's [45:16.490 --> 45:18.590] very, very expensive. [45:18.710 --> 45:20.590] It kind of drains your battery. [45:21.310 --> 45:22.530] So the other thing you can do [45:22.530 --> 45:24.710] is when you're looking to third party [45:25.550 --> 45:27.450] cloud providers, [45:27.450 --> 45:28.690] you can be really deliberate [45:28.690 --> 45:30.610] about asking them about their data [45:30.610 --> 45:32.840] retention, storage, [45:33.970 --> 45:35.130] disposal, [45:35.130 --> 45:36.710] all of their data policies [45:36.710 --> 45:38.890] because what your third party data provider [45:38.890 --> 45:39.410] does [45:40.870 --> 45:43.230] is basically what you do. [45:43.310 --> 45:45.850] And being a savvy customer. [45:45.970 --> 45:46.890] So pushing back [45:46.890 --> 45:49.010] if there aren't privacy protective services [45:49.010 --> 45:50.750] like you feel meets the threshold [45:50.750 --> 45:51.570] that you want [45:52.830 --> 45:55.110] to protect your users. [45:56.090 --> 45:56.810] And that's [45:56.810 --> 45:59.410] excellent because the change starts with calling it out. [45:59.690 --> 46:00.970] And bringing it [46:00.970 --> 46:02.770] to the attention of and paying attention to [46:02.770 --> 46:04.690] because you also think of all the different users [46:05.290 --> 46:06.750] and whether they're going [46:06.750 --> 46:07.970] to understand [46:09.730 --> 46:11.050] a 12 year old child [46:11.050 --> 46:13.090] is not going to care [46:14.110 --> 46:14.550] about [46:14.550 --> 46:16.370] all their different things, but it's like [46:16.370 --> 46:18.510] all they know is there's a lag time. [46:18.810 --> 46:20.330] And they're annoyed. It's like, no, let's [46:20.330 --> 46:22.950] make sure that we are paying attention, [46:22.950 --> 46:24.590] that we are bringing it in, and that we are [46:24.590 --> 46:26.570] reporting it, raising it. [46:26.570 --> 46:28.450] It's kind of the name and shame. [46:28.970 --> 46:30.630] Hey, this is a great game that you've [46:30.630 --> 46:32.270] created for children or geared [46:32.270 --> 46:34.310] towards this market. [46:34.390 --> 46:36.430] Now you're creating their permanent record [46:36.430 --> 46:37.810] for their school. [46:38.110 --> 46:40.630] The equivalent of that example [46:40.630 --> 46:42.630] that's going to follow them forever [46:42.630 --> 46:44.110] and is going to be based on [46:44.110 --> 46:46.230] AI modeling and [46:46.230 --> 46:48.250] data points that are not [46:48.250 --> 46:50.490] necessarily accurate. I mean, they're going to be based [46:50.490 --> 46:52.690] on, if we're looking at the early sets, [46:53.670 --> 46:54.650] the 5, [46:54.650 --> 46:56.470] 11, 5, 10 [46:56.470 --> 46:57.310] male [46:59.670 --> 47:00.270] interactions [47:00.270 --> 47:01.490] and all of that [47:01.490 --> 47:02.770] different thing. [47:02.770 --> 47:04.890] If you had to think of [47:04.890 --> 47:07.450] one way to bring it [47:07.450 --> 47:09.350] home, of how you're [47:09.350 --> 47:11.310] dreaming of electric sheep, [47:11.910 --> 47:13.370] what are some good takeaways [47:13.370 --> 47:15.510] other than, I mean, I love what you were saying [47:15.510 --> 47:17.610] with the proactive [47:17.610 --> 47:18.970] steps. [47:19.030 --> 47:21.270] What's a good nugget? [47:21.570 --> 47:23.230] What's the quote? [47:25.790 --> 47:27.250] It's the hangover spot [47:27.250 --> 47:29.690] on Sunday morning. I don't know if I have a quote. [47:40.350 --> 47:41.450] Someone asked what is [47:41.450 --> 47:43.530] inspiring in this space, which is [47:43.630 --> 47:45.190] a great question to end on because [47:45.890 --> 47:47.370] it's all doom and gloom [47:47.370 --> 47:49.010] when I go to the dinner party. [47:49.750 --> 47:51.290] XR is magic. [47:52.510 --> 47:53.830] It's magic. [47:54.570 --> 47:55.590] It makes, it can [47:56.810 --> 47:57.650] it can make [47:57.650 --> 47:59.850] you feel [47:59.850 --> 48:01.830] like you are in a different body. [48:02.070 --> 48:03.410] You start to embody [48:03.410 --> 48:05.810] the characteristics of your avatar [48:05.810 --> 48:07.470] in a social environment [48:07.470 --> 48:09.530] within two minutes. [48:09.990 --> 48:11.390] So if I make myself an [48:11.390 --> 48:13.590] eight foot tall dragon, I'm going to dominate [48:13.590 --> 48:15.690] that space in two minutes. [48:16.450 --> 48:17.690] It takes two and a half [48:17.690 --> 48:19.610] minutes for me to put you in [48:19.610 --> 48:21.550] my lab context and give you a third [48:21.550 --> 48:23.570] arm and have you pop three balloons [48:23.570 --> 48:25.230] at the same time. [48:26.570 --> 48:27.110] Neuroplasticity [48:27.110 --> 48:29.770] is magic. XR is a [48:29.770 --> 48:32.550] means to access your neuroplasticity, [48:32.550 --> 48:34.370] your imagination is [48:34.370 --> 48:36.510] really the limit. And we've never had [48:36.630 --> 48:38.710] a technology like this before. [48:38.930 --> 48:41.210] So let's not fuck it up. [48:41.510 --> 48:42.890] That's my quote. [48:42.970 --> 48:44.330] I like that. [48:44.330 --> 48:46.530] Thank you so much and thank you to [48:46.530 --> 48:48.450] the XR village for [48:49.770 --> 48:50.410] bringing [48:50.410 --> 48:53.310] this topic, inviting us. [48:53.410 --> 48:54.630] Please go check it out [48:54.630 --> 48:56.450] if you haven't already. I'm [48:56.450 --> 48:58.650] Lawyer Liz on Twitter. Please feel [48:58.650 --> 49:00.710] free to engage. [49:00.710 --> 49:02.750] I'm Britton Heller. You can [49:02.750 --> 49:04.930] find me as Britton Heller on [49:04.930 --> 49:06.990] Twitter or on LinkedIn. [49:08.170 --> 49:08.910] Thank you [49:08.910 --> 49:09.930] so much.