[01:23.200 --> 01:27.400]  All right, welcome everyone to our four o'clock session.
[01:29.240 --> 01:33.300]  Just want to remind everyone, put your phones on stun or vibrate.
[01:34.060 --> 01:35.080]  Stay hydrated.
[01:35.700 --> 01:36.760]  It's hot today.
[01:36.980 --> 01:41.180]  And I know there's no bottle of water, so there's water fountains around the corner there.
[01:42.100 --> 01:49.280]  Again, if you have any issues, medical or otherwise, look for someone with a radio or a staff shirt.
[01:50.840 --> 02:00.200]  So, I have the pleasure of introducing Giulia Corona and  Alessandro Polidoro, and they have the most creative title of, I think, the whole day today.
[02:02.200 --> 02:06.880]  Porn platforms hate them for exposing their mischief with these two weird tricks.
[02:07.220 --> 02:07.620]  Welcome.
[02:09.640 --> 02:10.580]  Hello, everyone.
[02:11.020 --> 02:11.240]  Hello.
[02:12.620 --> 02:13.400]  Thank you.
[02:13.580 --> 02:15.100]  We are Alessandro and Giulia.
[02:15.100 --> 02:36.400]  Alessandro is a lawyer, I'm a communication designer and data analyst, and we are part of this European organization called Tracking Exposed, who does algorithmic auditing, algorithm accountability, and focus on digital rights.
[02:36.400 --> 02:49.360]  In this talk, we will present you our research on Pornhub, and we also investigate TikTok and YouTube currently.
[02:50.120 --> 02:53.820]  So, check it out at tracking.exposed.
[02:55.920 --> 03:05.700]  What we do, basically, we analyze platforms, not people.
[03:06.540 --> 03:12.960]  We want to collect data to understand the proprietary algorithms for public interest.
[03:12.960 --> 03:43.600]  In this example, we use the Pornhub Tracking Exposed browser extension, and the idea is that platforms can conceptualize gender and other social constructs, and can verify conceptions about it, and can shape, affect, maintain, maintain, status quo, and gender identities.
[03:44.820 --> 03:47.900]  So, why porn, and why Pornhub?
[03:48.060 --> 04:11.020]  Porn, because we can consider Internet as a pornographic technology, if we think about food porn, or how they are made now to, let's say, get your attention visually, and stimulate your gaze.
[04:11.880 --> 04:12.800]  Why Pornhub?
[04:12.940 --> 04:21.560]  Pornhub, because it's the first and the biggest of the pornographic platforms, yet, at the moment.
[04:23.000 --> 04:27.980]  It is a platform because, first, there is identification of the user.
[04:27.980 --> 04:36.800]  They have content and there is standardized based on popularity and predictability.
[04:37.100 --> 04:41.660]  And they also have targeted content and ads.
[04:43.020 --> 04:46.840]  Third, they have this data porn.
[04:46.840 --> 05:10.180]  So, they use their blog, Year in Review, as a communication strategy and form of advertising, because they claim this is a transparency tool, but actually, they decide what to tell you, and what data they can share, and what data they cannot.
[05:12.080 --> 05:27.160]  So, we know that algorithms are black boxes, and there is no knowledge about how fair and unbiased algorithm should look like.
[05:27.360 --> 05:29.140]  We don't even know if it is possible.
[05:29.140 --> 05:37.340]  We know that automated surveillance is something related with the conservatives.
[05:38.740 --> 05:44.500]  But now, we give it for granted, and we just think that it's biased.
[05:45.140 --> 05:49.420]  And there is a universal truth, but we are not sure about it.
[05:49.420 --> 06:02.020]  And we cannot think about a different world, a different alternative to this system, and to surveillance, capitalism.
[06:03.160 --> 06:06.400]  So, what we do, we do scrapism.
[06:06.760 --> 06:12.920]  It's web scraping, but for artistic, emotional or critical ends, this is what Sam Lavinia says.
[06:13.160 --> 06:15.500]  So, we find ourselves in this.
[06:15.500 --> 06:28.380]  We have HTML websites and pages with non-structured data all nested around, and we created this extension to extract, parse them, and do scraping.
[06:28.700 --> 06:38.320]  Then we have a structured JSON or CSV that we can analyze and compare to also document what's happening with the platform.
[06:40.120 --> 06:49.740]  Our methodology for this paper was actually considering one sexual interest over the three PornApp allows us to have.
[06:50.000 --> 06:54.140]  So, we can be interested in girls, guys, or girls and guys.
[06:54.400 --> 06:58.380]  But then we can have different gender identities.
[06:58.880 --> 07:02.700]  This is all in the registration process.
[07:02.700 --> 07:11.340]  So, there are ten gender identities, and we decided to focus on ten accounts.
[07:12.220 --> 07:19.200]  We got, like, 1,600 on pages, and there were, like, 46 videos per page.
[07:19.460 --> 07:29.060]  But we understood that the unique videos were far away less, 118 for this study.
[07:30.280 --> 07:31.760]  How does it look like?
[07:31.880 --> 07:36.360]  This is one video, one entry for one video on the home page.
[07:36.660 --> 07:38.100]  So, we have different metadata.
[07:39.040 --> 07:43.340]  Title, author name, author link, and the size, whatever.
[07:43.720 --> 07:49.220]  We focused on section name because we wanted to analyze the home page from a section point of view.
[07:49.220 --> 07:55.560]  Because there were, there are different sections, as you may know, or not.
[07:55.760 --> 07:56.100]  I don't know.
[07:56.620 --> 08:05.300]  Now I will show you how personalization can look like from a visualization point of view.
[08:06.000 --> 08:09.500]  And the graphs were made with the open source tool, Gephi.
[08:10.080 --> 08:11.360]  So, check it out.
[08:11.360 --> 08:17.420]  And this is how content that is not personalized looks like.
[08:18.060 --> 08:27.300]  We have the ten triangles here that are our accounts we created.
[08:27.660 --> 08:31.660]  And the videos in the middle, so shared by everyone.
[08:31.660 --> 08:36.580]  And this was the case for hot porn videos in your country.
[08:36.920 --> 08:39.060]  Most viewed videos in your country.
[08:39.360 --> 08:41.600]  So, geographical personalization.
[08:42.620 --> 08:46.380]  And our bots were all from the same country.
[08:46.660 --> 08:48.580]  So, that's why you see it like this.
[08:48.740 --> 08:52.260]  And recently featured videos as well.
[08:52.980 --> 08:59.020]  But what happens when you give a look to another section, like recommended category for you?
[08:59.020 --> 09:06.160]  There is different, and it can also not be present for different gender identities.
[09:06.520 --> 09:11.340]  So, we just got it in six accounts.
[09:12.060 --> 09:19.800]  And we can start to see that there are proto-clusters.
[09:20.200 --> 09:26.700]  So, some accounts that share a lot of videos and others that don't.
[09:29.120 --> 09:39.580]  In the last individualized personalization in the recommended for you category, for you section, sorry.
[09:40.600 --> 09:43.680]  We can see two different clusters.
[09:44.020 --> 09:46.540]  And they are really well defined.
[09:46.820 --> 09:53.680]  So, we found out that there was a heteronormative cluster.
[09:53.680 --> 09:55.860]  So, male and female together.
[09:57.020 --> 10:02.340]  In a more standard cluster.
[10:02.740 --> 10:04.820]  Because they were sharing like a lot.
[10:05.380 --> 10:07.180]  And then there was another cluster.
[10:07.540 --> 10:13.200]  But always depending on the male one.
[10:14.780 --> 10:18.200]  There is a less heteronormative.
[10:19.260 --> 10:20.560]  But it puts all together.
[10:21.660 --> 10:26.680]  And it is the case of trans men, non-binary, same-sex couple.
[10:26.980 --> 10:27.620]  So, female.
[10:28.040 --> 10:28.520]  So, lesbian.
[10:29.320 --> 10:30.340]  And trans woman.
[10:32.720 --> 10:41.720]  The curious thing is that in the gender normative, we had the models, channel and porn stars.
[10:42.220 --> 10:47.680]  But in the other, we were not able to find channels.
[10:47.680 --> 10:55.800]  And so, production companies that are Pornhub's property as well.
[10:56.000 --> 10:59.960]  So, this was a little bit tricky.
[11:00.220 --> 11:00.760]  We don't...
[11:00.760 --> 11:04.200]  We still don't know how to give a meaning to this.
[11:06.640 --> 11:19.300]  But basically, our main argument is that the platform and the algorithmic suggestions were fostering a heteronormative perspective on sexual desire.
[11:19.540 --> 11:25.720]  And they were typical of heterosexual, white and hegemonic masculinity.
[11:27.200 --> 11:30.220]  So, what comes next to this?
[11:31.000 --> 11:36.040]  We would like to raise awareness about topics like the GDPR, like we are doing here.
[11:36.260 --> 11:36.540]  I don't know.
[11:38.120 --> 11:40.900]  Algorithms, consent, and especially desire.
[11:41.580 --> 11:52.360]  We would like to hold more the platform accountable for the lack of transparency with the respect of the tracking.
[11:52.360 --> 12:04.660]  So, we also decided to proceed in a different way and to give a better look at the cache and the cookies.
[12:05.520 --> 12:10.820]  And then, we would like to do some more call for actions through our tools.
[12:11.700 --> 12:17.800]  So, the next step was actually to create this new software called Guardoni.
[12:17.800 --> 12:32.900]  Because we understood that getting data from volunteers is great to have an empirical first example of what's happening and some anecdotal evidences.
[12:32.900 --> 12:44.040]  But then, to prove them and to say, okay, we saw them, we need to have a clean environment and a sort of clean setup.
[12:45.020 --> 12:49.400]  Even a small browser history can buy us everything.
[12:50.560 --> 12:53.040]  So, what are we now looking for?
[12:53.360 --> 12:59.740]  We are looking for third-party trackers and how they relate to recommended content.
[12:59.980 --> 13:01.420]  And basically, okay, born.
[13:01.660 --> 13:03.800]  And how it works.
[13:04.040 --> 13:13.760]  How the rabbit hole can make people become more sadistic or stuff like that.
[13:15.520 --> 13:19.660]  So, this tool is available on GitHub.
[13:20.440 --> 13:26.780]  And you can find it in the methodology directory.
[13:28.300 --> 13:29.940]  What can it do for us?
[13:30.020 --> 13:32.180]  So, it's a laboratory setup.
[13:33.360 --> 13:39.740]  It can help us installing the extension on a clean browser because it's directly installed.
[13:41.920 --> 13:51.620]  Initialize a new Chromium, in this case, profile, a new user data directory.
[13:52.460 --> 14:02.820]  It can automatically navigate through the website and gain some tracking data and screenshots.
[14:03.940 --> 14:07.480]  We can have a look at the code.
[14:07.900 --> 14:12.240]  It's basically a node library with JavaScript extensions.
[14:13.340 --> 14:18.960]  And the APIs are quite simple.
[14:19.320 --> 14:28.900]  You can launch the browser, wait, go to a precise URL, take a screenshot and then close it.
[14:28.900 --> 14:30.620]  This is a simple example.
[14:31.180 --> 14:38.560]  But an experiment that we do can be exactly looking like this.
[14:41.820 --> 14:44.560]  The experiment is a set of comments.
[14:45.220 --> 14:51.060]  We create a JSON with a list of URL we want to access.
[14:51.500 --> 14:57.040]  And for how much time, if we want a screenshot, it's all defined in a JSON.
[14:57.040 --> 15:00.820]  You can create it with a GIST on GitHub.
[15:03.060 --> 15:13.540]  And when you give it to Guardoni, Guardoni automatically navigates through these comments and instructions.
[15:14.680 --> 15:19.980]  The other thing we are doing is to collect the cookies in the local storage.
[15:20.220 --> 15:23.620]  Local storage is a cache, again, on your browser.
[15:23.620 --> 15:30.380]  It is different from web storage of cookies because it's permanent.
[15:32.880 --> 15:35.280]  The cookies don't expire.
[15:36.560 --> 15:40.700]  So, as you can see, we collect the cookies.
[15:40.700 --> 15:52.440]  But also, we are creating a sort of white list of the domain Pornhub.com and Pornhub.ncdn.com.
[15:52.680 --> 15:56.500]  So, we know it's their cookies and okay.
[15:57.020 --> 16:04.640]  But then we are also extracting the third-party trackers and parse the domain.
[16:04.640 --> 16:09.160]  So, we can know who is behind these cookies.
[16:09.480 --> 16:23.960]  Because in the porn industry, online there was a bubble and no one wanted to advertise on porn platforms.
[16:23.960 --> 16:36.380]  So, basically, there is a different market and a parallel ecosystem of third-party service providers.
[16:36.680 --> 16:41.580]  And we would like to know more because it has not been scrutinized yet.
[16:41.580 --> 16:49.040]  But by regulators or research community, academics and policy makers.
[16:50.000 --> 16:55.720]  And now, Alessandro will talk to you more about the legal perspective.
[16:57.620 --> 16:58.920]  Thank you.
[16:58.920 --> 16:59.440]  Thank you, Julia.
[17:08.640 --> 17:09.440]  Sorry.
[17:10.160 --> 17:10.280]  Okay.
[17:10.280 --> 17:11.920]  Can you hear me?
[17:12.160 --> 17:12.440]  All right.
[17:13.940 --> 17:14.200]  So, yeah.
[17:14.880 --> 17:15.360]  Hello.
[17:15.620 --> 17:16.960]  Thank you for being here.
[17:17.180 --> 17:22.980]  Thank you, Julia, for the presentation of our tools and the research we have done with Tracking Exposed in the past years.
[17:23.600 --> 17:24.460]  So, yeah.
[17:27.880 --> 17:34.840]  Basically, my role here as a lawyer is to try and answer to the following questions.
[17:35.080 --> 17:36.340]  So, how did we get here?
[17:36.340 --> 17:49.400]  How did it happen that such a huge and invasive market that deals with an incredibly sensitive aspect of our culture, our society...
[17:50.260 --> 18:05.540]  How did it happen that this market is basically growing unregulated, unsupervised, and sometimes even just being a mockery of the principles that should inform the digital sphere.
[18:05.860 --> 18:35.400]  We've seen that, especially when we investigated PornHub, we realized that on one end we have an official and formal statement of acknowledging different gender orientation, different sexual preferences, with the nine possible categories of sexual orientation that a user is allowed to express to the platform before navigating it.
[18:35.580 --> 18:48.540]  But, in fact, the algorithm is biased in the sense that it basically gives you either straight heteronormative content recommendation or the weird content recommendation.
[18:49.720 --> 18:50.920]  Is this even legal?
[18:51.780 --> 18:53.100]  How does it work, right?
[18:53.240 --> 18:57.700]  This is the starting point of what I'd like to discuss with you today.
[18:59.380 --> 19:02.520]  And the answer is not that simple.
[19:02.840 --> 19:12.100]  And this is especially because of this word that you see right here, this legal vacuum concept that may be not too familiar for most of you.
[19:12.100 --> 19:18.440]  For us legal people dealing with the digital world, it's actually the norm.
[19:18.620 --> 19:19.960]  It's our everyday life.
[19:20.160 --> 19:23.320]  Is this a legal context in which things are not clear?
[19:23.780 --> 19:28.160]  There is no applicable law or in which some injustice is uncorrected.
[19:28.700 --> 19:33.820]  This has been for a very long time, as I said, the general norm for the digital world.
[19:33.820 --> 19:42.660]  We are slowly trying to fill up the void to put new solutions, new ideas into this vacuum.
[19:43.440 --> 19:49.760]  Of course, I can speak for the European legal context.
[19:50.060 --> 19:52.320]  And many, many step forwards have been done.
[19:52.460 --> 19:53.640]  But in the U.S., it's the same.
[19:53.640 --> 20:04.700]  I think, for instance, all the new acts that came out in California with the California Consumer Privacy Act and the California Privacy Rights Act, if I'm not mistaken.
[20:06.540 --> 20:07.240]  So, yeah.
[20:07.800 --> 20:09.040]  We are slowly.
[20:09.420 --> 20:15.860]  And slowly has a double meaning in this context because it's the usual slowness of the legal world.
[20:16.040 --> 20:22.660]  And also, the fact that the digital world is, by nature, very, very fast compared to the other things of the world.
[20:22.660 --> 20:23.960]  We see it every day.
[20:24.360 --> 20:31.480]  And so, when you think of the fact, for instance, Julia mentioned GDPR, which is the General Data Protection Regulation.
[20:32.300 --> 20:34.580]  It's enforced in Europe since 2018.
[20:34.740 --> 20:39.300]  Well, it took 20 years to come up with that.
[20:39.480 --> 20:44.160]  Because the law that was existing before that was from 1995.
[20:46.480 --> 20:47.440]  And, yeah.
[20:47.660 --> 20:48.100]  You see.
[20:48.880 --> 20:51.240]  I will talk more about it in the upcoming slides.
[20:51.240 --> 21:05.560]  But what I'm trying to do here is to give you the feeling of the different pace, different speed that you can experience in the digital world and in the legal world compared.
[21:06.120 --> 21:07.820]  So, about this legal vacuum.
[21:08.180 --> 21:11.220]  Legal vacuum is a catch-all term.
[21:11.380 --> 21:17.280]  We can actually divide it a little bit more and try to understand it a little bit more.
[21:17.440 --> 21:20.160]  Of course, it's the lack of norms in general.
[21:20.720 --> 21:21.940]  That's quite intuitive.
[21:22.280 --> 21:27.560]  But in this case, we can think of it at every level, every kind of norm.
[21:27.560 --> 21:34.460]  So, it can either be actual law, like international conventions or national law, you name it.
[21:34.900 --> 21:37.780]  Or also other kind of rules, other kind of norms.
[21:38.040 --> 21:40.660]  Let's think of internal policies of these platforms.
[21:40.940 --> 21:56.500]  I mean, maybe we don't even have to talk about these problems if only the pornographic platforms were to have decent policies of content moderation and protection of personal data of their users.
[21:56.500 --> 22:01.120]  But here we are, so apparently, something didn't work out as planned.
[22:01.760 --> 22:06.880]  And the problem is that that's not even the only thing that we are lacking right here.
[22:07.060 --> 22:13.440]  Because lack of concept is also a huge component of this legal vacuum.
[22:14.800 --> 22:18.500]  Because you see, there is...
[22:20.080 --> 22:22.720]  So, lawmakers are basically people.
[22:22.720 --> 22:24.000]  And we are people too.
[22:24.120 --> 22:27.340]  So, we can somehow imagine to be lawmakers ourselves.
[22:27.680 --> 22:38.860]  And if we try and imagine ourselves having to create new rules for the digital world, especially when it comes to the pornographic world, we actually don't have categories.
[22:39.140 --> 22:44.280]  We would immediately meet a lack of principles.
[22:44.280 --> 22:46.360]  Think of revenge porn.
[22:47.120 --> 22:51.400]  You may have heard about this very problematic practice that is happening.
[22:51.860 --> 22:56.960]  The proper term for it would be non-disclosure dissemination of intimate contents.
[22:57.160 --> 23:06.240]  Basically, it's when somebody is sharing a naked picture of somebody else without his or their consent.
[23:06.240 --> 23:17.620]  And it took so long to the legal community to start acknowledging the concept of actually the intimate material.
[23:17.880 --> 23:29.980]  The idea that a photo, a video can actually be not only just a good that you trade and exchange, a source of entertainment, but it can also be a mean of sexual expression of an individual.
[23:31.820 --> 23:38.800]  And actually, sexual expression is another principle that is not really clear in the legal world.
[23:38.920 --> 23:40.320]  Think of that.
[23:40.420 --> 23:41.660]  Think of the gender expression.
[23:41.940 --> 23:43.980]  I mean, it's a huge topic right now.
[23:43.980 --> 23:50.600]  And, of course, everybody should have the freedom to express themselves.
[23:51.060 --> 24:03.500]  There is the right to self-determination that appears here and there everywhere, especially in the Western judicial cultures.
[24:03.500 --> 24:12.560]  But we don't really have a legal theory that encompasses all the aspects of what we could understand as gender expression.
[24:12.560 --> 24:15.560]  And so this makes it even more difficult.
[24:15.900 --> 24:24.960]  And, of course, as if this was not enough already, there is also a serious lack of political will to tackle these problems.
[24:25.240 --> 24:34.160]  To tackle the legal vacuum, especially in the pornographic sphere, online pornography, but also in the digital world in general.
[24:34.160 --> 24:41.680]  Right now, this last part that I just said is changing because of, of course, how catchy all the digital topics are.
[24:42.200 --> 24:46.620]  But imagine, again, put yourself in the lawmakers' dress, right?
[24:46.800 --> 24:48.140]  In their shoes.
[24:48.140 --> 24:54.240]  And try to imagine going to your electoral base to do the people that support you.
[24:54.360 --> 25:02.960]  And say that tomorrow you want to dedicate your work and your time in protecting the right of, I don't know, online porno models.
[25:03.220 --> 25:09.040]  Or people who want to just access these porn platforms, right?
[25:09.040 --> 25:15.340]  It's not exactly the most appealing thing given also the cultural taboo.
[25:15.640 --> 25:17.880]  We will get back on it very, very, very soon.
[25:21.880 --> 25:33.560]  So, questions like, I don't know, like, is it convenient for a political figure to address these kind of topics can actually be a slippery slope.
[25:33.740 --> 25:42.000]  And at least, for sure, it's one of the main problems that is making the filling up of this vacuum very, very difficult.
[25:42.000 --> 26:00.080]  And now, you see, at the bottom of this slide, I am quoting Ludwig van der Rohe, the famous architect, with this very famous sentence, less is more, because we may think that actually, sorry, we may think that actually less rules, less norms, especially,
[26:00.340 --> 26:07.880]  I guess, the most libertarian anarchist people in the room today may probably agree with this.
[26:07.880 --> 26:11.980]  Not always too many rules means less problems.
[26:14.280 --> 26:17.360]  But is it the case for us right now?
[26:18.380 --> 26:22.120]  So, let's talk about the European context.
[26:22.620 --> 26:25.540]  Again, this is probably a bit far away from...
[26:25.540 --> 26:30.040]  I don't know, is there any person from Europe or there is not from the U.S. in the room?
[26:30.540 --> 26:31.100]  Oh, okay.
[26:32.920 --> 26:34.120]  Okay, so...
[26:34.120 --> 26:36.000]  But don't worry, American friends.
[26:36.000 --> 26:40.420]  This can actually be an interesting perspective for you as well.
[26:42.160 --> 26:47.620]  So, you see, I mentioned GDPR, you see it in the first line, General Data Protection Regulation.
[26:48.460 --> 26:57.560]  This has been a huge thing for us in Europe, but not only for us, because the GDPR applies to any subject who is running business in the EU.
[26:57.560 --> 27:10.440]  So, even if you are not a European citizen, but you want to run business in the EU, you still need to take care of this and be careful not to violate this rule, because the fines are very, very high.
[27:10.440 --> 27:31.500]  And this rule, this regulation specifically, has the merit to have introduced a beautiful concept that was discussed for a very long time, but it took very long to appear in a legal text, which is the concept of privacy by design, as opposed to privacy by default.
[27:31.500 --> 27:45.400]  So that, at some point, platforms had to realize that to ensure privacy of their users was not enough to just promise that they will do it.
[27:45.900 --> 27:48.380]  It's not enough to just say, don't worry, we will take care of it.
[27:48.440 --> 27:59.420]  You have to prove that the design of your platform, the cybersecurity measures that you put in place, the whole architecture of your algorithms, is actually privacy-oriented.
[27:59.780 --> 28:04.820]  That was a huge step forward for the data protection law community.
[28:05.020 --> 28:17.120]  And it only happened four years ago, let's say six, because the regulation was approved in 2016, but still, it's, I mean, computers have been around for a bit longer than that.
[28:18.700 --> 28:26.600]  Again, going back in time, we have the privacy, e-privacy directly, the privacy and electronic commerce directive that you see since 2022.
[28:27.900 --> 28:39.120]  That was actually also very important, because it introduced concepts like data retention, the limits to data retention, and also was one of the first attempts in Europe to regulate cookies.
[28:39.120 --> 28:43.460]  That was a huge topic back then, it is still now.
[28:43.980 --> 28:45.840]  Or also the e-commerce directive.
[28:46.000 --> 28:52.100]  The e-commerce directive has the merit of introducing one of the first attempts to categorize digital platforms.
[28:52.340 --> 29:01.020]  In the text of the norm, they are called intermediary service providers, or ISP, also called Internet service providers, if you like.
[29:01.020 --> 29:05.560]  And they introduced a brilliant idea of dividing them in three categories.
[29:05.880 --> 29:06.660]  Three categories.
[29:07.340 --> 29:14.600]  One, well, not that it matters, but one is mere conduit, one is cash provider, and one is hosting provider.
[29:14.840 --> 29:34.380]  Funny thing about this categorization is that eventually the last one, the hosting provider category, have become the main one, which is good for many reasons, because it's actually a gradual increase of norms and regulation coming from the first one that I mentioned,
[29:34.540 --> 29:38.460]  the mere conduit, going through the cash, and then the hosting provider.
[29:38.580 --> 29:42.220]  So hosting providers have to abide to more rules than the others that I've mentioned.
[29:42.400 --> 30:02.500]  And right now, maybe some of you have heard about it, that the Digital Service Act and the Digital Market Act in Europe, these two new regulations that are about to be put in place, a couple of years, hopefully, will also implement even more this concept.
[30:02.660 --> 30:13.120]  So I'm trying to give you some very basic, not even legal tools, but just juridical ideas to play with.
[30:14.100 --> 30:29.940]  Because right now, we will start now diving into any possible legal solutions, or at least a legally informed perspective to this pornographic world, the online pornography world that we're trying to understand a little bit better.
[30:30.520 --> 30:34.040]  So, legally speaking, what about porn platforms?
[30:35.200 --> 30:43.400]  The first sentence that I decided to mention here in the slide is from Potter Stewart, American legal scholar, judge and lawyer.
[30:44.240 --> 30:48.320]  Well, this sentence is out of context, but it was during the trial.
[30:48.500 --> 30:57.380]  He was trying to express what a pornographic content is, what a porn material is.
[30:57.480 --> 30:59.020]  I'm sorry for the loss in translations.
[30:59.320 --> 31:01.000]  We are not English native speakers.
[31:01.000 --> 31:02.580]  Sometimes I can get clumsy.
[31:03.060 --> 31:06.240]  Anyhow, he was talking about this pornographic image.
[31:06.440 --> 31:13.500]  And he said, I shall not today attempt further to define this kind of material, but I know when I see it.
[31:13.900 --> 31:16.000]  So, you see, when he used this sentence...
[31:16.700 --> 31:17.900]  Well, it was long ago.
[31:18.080 --> 31:20.960]  I think in the 60s or even late 50s.
[31:20.960 --> 31:32.380]  Anyhow, when he used this sentence, he gave the perfect example of what's the difficulty of wanting to talk about something that we can't even define.
[31:32.760 --> 31:35.740]  And even more than that, we don't dare to define.
[31:36.080 --> 31:38.140]  Because of the cultural taboo.
[31:38.480 --> 31:40.140]  Again, we will get to it.
[31:40.300 --> 31:42.060]  Just trying to plant a seed right now.
[31:43.440 --> 31:46.080]  But yeah, we still have this problem nowadays.
[31:46.080 --> 31:56.380]  It's still a way to giggle, to think that it's a bit of a naughty thing to talk about pornography, right?
[31:56.520 --> 32:03.300]  Even though it's in the everyday life of the majority of the people we know, especially the younger generations.
[32:04.080 --> 32:08.580]  And we can already see the consequences of that.
[32:08.580 --> 32:11.260]  So, the idea is to...
[32:12.300 --> 32:21.660]  The reason why I decided to start with this quote is because I think that there is a clear mistake that we should surpass.
[32:22.140 --> 32:27.540]  We should try to dare and define what we're talking about when we talk about porn.
[32:29.080 --> 32:33.180]  Producing, selling, and possessing porno is still illegal in many nations.
[32:33.180 --> 32:34.820]  This is funny, right?
[32:35.120 --> 32:36.680]  Because it raises a question.
[32:37.060 --> 32:45.500]  When we want to find legal solutions or any kind of solutions to the pornographic world, we still have to worry if...
[32:45.500 --> 32:47.860]  I mean, is this a first word problem?
[32:48.700 --> 32:53.620]  If you allow me this clearly joking expression.
[32:55.520 --> 32:56.800]  I don't think so.
[32:57.120 --> 32:58.920]  But in a certain sense it is.
[32:58.920 --> 33:17.180]  For sure, when we look at the data, when we look at the map of the globe, we see that porno is illegal in most of the countries that don't belong to the western block, the western quarter of the planet.
[33:19.100 --> 33:34.060]  Nevertheless, we can't deny that the industry affects the whole planet, especially considering that many of the actors, voluntary or involuntary actors that end up on these platforms, come especially from the other parts of the world.
[33:34.060 --> 33:43.300]  So, the problem is global, but there is probably a clear power dynamic going on.
[33:43.540 --> 33:55.680]  This is also a way for us to be a bit more aware of the colonial implications of yet another part of the world and of the digital world.
[33:55.680 --> 34:01.300]  And there has been a significant effort to stop child pornography and non-consensual pornography.
[34:01.300 --> 34:20.940]  So, the reason why I mentioned that here in the third point is because after making you feel guilty and making you feel a bit worried probably about how big this issue is, how big the problem that we're trying to solve is, there are actually already some very relevant and actual effective solutions in place.
[34:21.080 --> 34:28.780]  They are all very narrow in their focus, given the huge dimension of the problem at hand.
[34:29.300 --> 34:34.900]  But yet, they are focused and partial to this specific topic, but yet very important.
[34:35.160 --> 34:39.740]  I think, as I was mentioning non-consensual pornography, but child pornography is also very obvious.
[34:39.740 --> 34:46.820]  In these cases, of course, it's easier for a lawmaker to imagine a solution.
[34:46.980 --> 34:52.820]  First off, because creativity, apparently they say, works better when it's limited in a specific context.
[34:52.980 --> 35:10.800]  And also because it's not only the sexual expression of those who make porno or consume porno, but it's also a matter of protecting the rights of other people that may get hurt in the transaction, in this dynamic.
[35:11.160 --> 35:18.660]  So, what we can see, we can see a growing effort in preventing underage users to access online pornography.
[35:19.320 --> 35:23.940]  This is another thing that, for instance, right now in the UK...
[35:23.940 --> 35:27.660]  Well, right now, a couple of years back started, maybe a little bit more.
[35:27.880 --> 35:32.220]  But also in other places of the world, that's a topic that is interesting.
[35:32.500 --> 36:01.000]  Because, I mean, of course, we want to imagine a safer way to protect the underage, the most sensitive people, those who are still maturing and creating their own sexual persona, from an unlimited and reckless access to these sometimes even difficult to digest and understand kind of contents.
[36:01.240 --> 36:09.080]  But on the other hand, think of the solutions that have come out when thinking of how to prevent underage users to access online pornography.
[36:09.080 --> 36:15.080]  The idea of requiring an ID when you connect on the Internet, that's even worse.
[36:16.080 --> 36:24.640]  Of course, the problem of minors watching porn can be very upsetting.
[36:24.920 --> 36:31.880]  But we always have to be careful when we think of solving a problem, not to solve it with an even bigger problem, right?
[36:32.120 --> 36:42.380]  Because I don't think that anybody here would be happy with a computer that asks you to show your driver's license before you log into whatever you want to log into that day.
[36:42.380 --> 36:49.040]  And, yeah, an interesting debate has also sparked over the addictive nature of mainstream pornography content.
[36:49.300 --> 36:51.520]  That's also very interesting and recent.
[36:51.760 --> 36:54.000]  And again, talking about solutions...
[36:56.720 --> 37:01.580]  Well, imagine a world where, like for cigarettes or...
[37:01.580 --> 37:05.980]  Well, usually in a... I mean, some of these...
[37:05.980 --> 37:07.500]  Cigarettes is a good example, actually.
[37:07.660 --> 37:11.720]  Imagine a world where you have to show a banner that says, Porn is addictive.
[37:12.000 --> 37:20.840]  Every time you access any video or any content that may involve explicit display of a human body.
[37:20.920 --> 37:21.960]  Sexual display of a human body.
[37:22.680 --> 37:24.460]  That's another kind of solution.
[37:24.460 --> 37:26.680]  That can be... That's a world we don't live in.
[37:27.060 --> 37:28.220]  That we can imagine.
[37:29.140 --> 37:29.700]  And...
[37:29.700 --> 37:31.880]  But then again, it's...
[37:31.880 --> 37:33.460]  It can also be...
[37:34.160 --> 37:45.300]  Specifically, when recognizing the addictive nature of this kind of contents, it can also be yet another way to create dystopian and creepy solutions.
[37:45.580 --> 37:49.860]  Think of all those who just would like to ban pornography altogether from the Internet.
[37:49.860 --> 37:50.860]  Because it's obscene.
[37:50.980 --> 37:51.620]  Because it's wrong.
[37:51.840 --> 37:53.680]  Because it's not...
[37:54.520 --> 37:55.280]  It's not clean.
[37:55.440 --> 37:55.980]  It's dirty, right?
[37:56.140 --> 37:57.980]  And so people would just like to suppress that.
[37:58.080 --> 38:01.760]  And think of all the people that work in the industry and maybe have this as the only source of income.
[38:02.000 --> 38:04.320]  Think of all the people that just like to do it.
[38:04.480 --> 38:06.020]  Can we really make that illegal?
[38:06.300 --> 38:15.080]  This is a question that I'm raising because we will carry it with us to the next part and the final part of this conversation that we're having today.
[38:15.300 --> 38:18.440]  By the way, in case you want to interrupt, ask questions.
[38:18.440 --> 38:19.500]  I don't know.
[38:19.600 --> 38:20.240]  Raise a point.
[38:20.460 --> 38:20.780]  Feel free.
[38:20.920 --> 38:22.140]  There is a microphone over there.
[38:22.280 --> 38:23.060]  You can just raise your hand.
[38:23.160 --> 38:24.640]  Interrupt me whenever you want.
[38:25.440 --> 38:26.740]  And just to...
[38:26.740 --> 38:28.560]  I was just checking the vibe.
[38:28.840 --> 38:29.020]  Okay.
[38:29.860 --> 38:32.300]  Some open issues, of course.
[38:32.720 --> 38:45.780]  The issues that I just described are actually issues where I could, if you want, later on because we don't have time right now, point to some solutions somewhere in the world where legal people or lawmakers are trying to tackle these issues.
[38:45.780 --> 38:52.300]  But there are still some other open issues such as, this of course is not a complete list, sex worker and prostitutes rights.
[38:52.940 --> 39:04.600]  We still have to remember that, again, talking about Pornhub for instance, it's very recent, their decision to remove from their platform all the non-consensual material that was still there.
[39:04.780 --> 39:12.300]  And think of all the people that ended up without their will because they were unaware of it or also those who were forced to end up there.
[39:12.460 --> 39:13.500]  What's with these people?
[39:13.660 --> 39:14.920]  What's with the sex workers?
[39:15.700 --> 39:20.440]  Of course, this is yet another huge topic.
[39:21.680 --> 39:31.180]  And when we think about the rights of sex workers, we always have to remember, in my opinion, that most of them work mainly online nowadays.
[39:32.060 --> 39:36.240]  What about the privacy, both of these workers and the users?
[39:36.960 --> 39:40.340]  Specifically, this topic is very dear to me.
[39:41.020 --> 39:46.600]  Also, given the research that we have conducted with Tracking Exposed that Julia was telling you about earlier on.
[39:48.960 --> 39:51.720]  There is a lot of data going on in this market.
[39:52.180 --> 39:54.720]  I mean, this raises many questions.
[39:54.720 --> 40:00.820]  For instance, is my search history on a porn platform sensitive data?
[40:02.060 --> 40:05.180]  That's a question, but it sounds very legal like this.
[40:05.320 --> 40:07.320]  I can also rephrase it in a different way.
[40:07.460 --> 40:15.380]  Would you share your search history on your favorite porn platform with a friend or an acquaintance or anybody?
[40:15.680 --> 40:20.600]  Or would you actually think that that's part of your very intimate life?
[40:20.600 --> 40:21.380]  I don't know.
[40:21.680 --> 40:24.020]  I'm asking you these questions not because I know the answer.
[40:24.180 --> 40:27.580]  Actually, I don't think that anybody can know an answer that fits to all.
[40:27.760 --> 40:32.260]  But because we should start this conversation.
[40:32.800 --> 40:34.900]  We should try to ask ourselves.
[40:35.180 --> 40:48.440]  If we don't want other people, lawmakers, figure powers, power figures to decide for us, we should start asking ourselves, what's the solution that I would like to see in the world?
[40:48.440 --> 40:51.120]  And what's the role of pornography in our society?
[40:51.620 --> 40:52.520]  That's the question.
[40:53.280 --> 40:54.840]  I mean, is it just wrong?
[40:57.180 --> 41:00.880]  Is it maybe natural?
[41:01.200 --> 41:05.320]  Or, I don't know, is it a business that is still worth preserving?
[41:05.540 --> 41:05.880]  I don't know.
[41:06.040 --> 41:10.660]  These are all questions that we should ask ourselves a little bit more often, I think.
[41:11.900 --> 41:14.720]  Also, because you see, when we...
[41:14.720 --> 41:17.120]  This is the final part of the presentation, by the way.
[41:17.300 --> 41:23.120]  When we think of solutions, when we try to harvest...
[41:23.120 --> 41:24.740]  Do you want to ask a question?
[41:25.060 --> 41:25.760]  At some point.
[41:27.200 --> 41:27.400]  Sure.
[41:27.400 --> 41:35.340]  Do you actually know about the discoverability and retention period of search logs in Porthub's properties?
[41:37.060 --> 41:45.500]  I mean, I know that it's written in their privacy policy, but there is no way for us to know actually what's happening in their company's server.
[41:45.760 --> 41:47.720]  So, I think it's four years.
[41:48.800 --> 41:49.260]  Okay.
[41:50.100 --> 41:50.980]  Was it the right answer?
[41:50.980 --> 41:52.260]  I don't know.
[41:53.720 --> 42:01.600]  I mean, their transparency report mostly talks about their reports to the NCMEC and their content moderation policies.
[42:01.820 --> 42:02.160]  Hmm.
[42:02.780 --> 42:05.640]  And not their privacy policy, as far as I can tell.
[42:06.400 --> 42:12.400]  Yeah, I'm looking forward to see their transparency reports, especially when the DSA, the Digital Service Act, will come into force.
[42:12.400 --> 42:21.000]  Because there is an article that will force large online platforms to make yearly transparency reports, which not all of them do right now.
[42:21.500 --> 42:25.460]  And also, not only that, they will have to abide to very strict criteria.
[42:25.660 --> 42:28.280]  So, it will not be your own made transparency report.
[42:28.440 --> 42:31.160]  It will be much more, well, full of stuff.
[42:31.160 --> 42:31.560]  Hopefully.
[42:31.700 --> 42:32.160]  We will see.
[42:33.820 --> 42:35.800]  I don't know if I answered to your question.
[42:35.980 --> 42:37.040]  Well, I have a larger question.
[42:37.040 --> 42:43.080]  I fail to understand what your problem is with Pornhub's recommendation system.
[42:43.380 --> 42:52.780]  And, you know, what the expectations of Pornhub users are about what kinds of videos are suggested to them.
[42:52.960 --> 42:55.520]  You know, I built recommender systems.
[42:55.810 --> 43:14.920]  And I can understand how they might be better at doing the heteronormative recommendations and categorization than doing the kinky side of the house, because there's much less history, much less tagging, much less crowdsourced information and metadata about that kind of content.
[43:15.180 --> 43:27.980]  So, it wouldn't be that easy for them to categorize people's kinks, you know, especially when they're, you know, they might cluster in strange, you know, less traditional ways.
[43:28.520 --> 43:34.320]  So, I'm not sure what your problem was with them, to actually go back to the initial talk.
[43:34.600 --> 43:36.880]  I mean, I don't know if Julia wants to pick this one, yeah.
[43:37.160 --> 43:38.720]  Well, it's not just our problem.
[43:39.020 --> 43:42.860]  I mean, since the 80s, feminism has always...
[43:42.860 --> 43:45.600]  If you could take your mask off, it might be better.
[43:47.100 --> 43:48.600]  I forgot about that.
[43:49.740 --> 43:53.020]  I mean, it's not just our problem with that.
[43:54.080 --> 43:57.140]  It's a feminist problem since the 80s.
[43:57.440 --> 44:00.400]  There was... there were the war on pornography.
[44:01.180 --> 44:10.320]  So, the idea that these recommender systems can now radicalize people and has...
[44:10.320 --> 44:25.440]  I mean, if you see that Facebook or Instagram can shape your reality, why this platform shouldn't shape your sexuality.
[44:25.760 --> 44:29.200]  Because a lot of videos in Pornhub are abused.
[44:31.560 --> 44:36.820]  I mean, it's paid... filmed prostitution.
[44:37.780 --> 44:38.600]  So...
[44:38.600 --> 44:39.540]  So, I'm sorry.
[44:39.840 --> 44:47.740]  Were your findings that people's sexuality was being steered in a way that was inappropriate or wrong?
[44:47.740 --> 45:06.000]  We are not at that point yet because there is a lot of work that has to be done first and to identify how this can affect the population as a sociological problem.
[45:06.840 --> 45:15.000]  I mean, can be technical, the solution of a social problem, no?
[45:15.480 --> 45:16.000]  Don't you agree?
[45:16.580 --> 45:18.260]  If I may step in.
[45:19.520 --> 45:27.220]  Beside the fact that you perfectly described in your own question the main problem of it, right?
[45:27.280 --> 45:46.760]  You said we have less data for this outskirt fringe of sexualities and gender orientation, so our algorithms cannot be as precise as for the, let's say, mainstream basic sexual orientations, right?
[45:47.280 --> 45:53.560]  And wouldn't it be already enough to ask ourselves if we are fine with making this kind of algorithms?
[45:53.560 --> 46:08.760]  And just to answer your question, we don't have evidences of the actual change of sexual orientation or any other empirical measurement of the impact of this content recommendation.
[46:08.760 --> 46:12.320]  But if we want to keep it a bit more simple, it's quite intuitive.
[46:13.240 --> 46:34.180]  We can imagine that especially a young person that is accessing a platform without knowing how the algorithm works, without knowing how the content is recommended to them, well, these young people will just consume what they see and what they have.
[46:34.320 --> 46:42.480]  And what if the algorithm has decided that you are heterosexual or homosexual when actually inside of you you are still forming a consciousness about it?
[46:42.520 --> 46:43.660]  What's going to happen with you?
[46:43.840 --> 46:54.380]  And we could still say that it's their fault because they decided to do that, they decided to access the platform, they decided to expose themselves to this influence of the algorithm.
[46:55.320 --> 47:02.440]  We could say that or we could not and realize that this is still part of a societal problem.
[47:02.680 --> 47:06.540]  This is still part of something that we should not turn our head away from.
[47:06.840 --> 47:14.140]  It sounds like no algorithm can do a satisfactory job given the absence of expectations and norms.
[47:14.440 --> 47:16.360]  So I'm not sure...
[47:16.360 --> 47:18.440]  I think that many people would agree with you.
[47:18.440 --> 47:26.260]  I, for one, would, and I don't know, could it be a reason to reimagine these algorithms?
[47:26.560 --> 47:32.920]  To, I don't know, introduce a bit more of privacy by design or, I don't know, respect by design?
[47:33.040 --> 47:33.720]  I don't know.
[47:33.820 --> 47:36.040]  I don't know if I managed to answer your questions.
[47:36.040 --> 47:37.500]  But we can continue afterwards.
[47:37.660 --> 47:38.100]  Thank you very much.
[47:38.280 --> 47:41.360]  Actually, so that dovetails nicely into this online question.
[47:41.360 --> 47:49.540]  Traditionally, the porn industry has been the technology leader, pushing VHS over Betamax, DVD protocol selection, streaming technologies.
[47:49.920 --> 47:54.280]  Do you see them leading in solving this bias algorithm?
[47:54.440 --> 47:56.460]  I'm sorry, I didn't quite grasp the...
[47:56.900 --> 48:07.420]  So, the porn industry has traditionally been the leader in technology selection for our media formats.
[48:08.080 --> 48:13.820]  Do you see them being a leader in technology in building and solving this bias issue?
[48:14.440 --> 48:18.840]  That was actually part of the final slides that I wanted to show.
[48:20.020 --> 48:22.480]  I will try to answer this question also.
[48:22.780 --> 48:24.440]  I don't know how much time we have left.
[48:24.660 --> 48:25.260]  Five minutes?
[48:25.460 --> 48:26.260]  Okay, let's try.
[48:28.020 --> 48:28.760]  So, yeah.
[48:29.120 --> 48:32.440]  Because the question was about solutions and my final slides also.
[48:32.720 --> 48:33.960]  So, what a nice coincidence.
[48:36.180 --> 48:39.960]  As you see here, possible approaches, possible solutions.
[48:39.980 --> 48:41.440]  I think that there are at least three.
[48:41.580 --> 48:42.300]  One is policy.
[48:42.580 --> 48:43.380]  One is advocacy.
[48:43.700 --> 48:46.340]  And one is litigation that I kept for last.
[48:47.360 --> 48:48.200]  So, policy.
[48:48.860 --> 48:52.940]  Currently, most of the power and agency over this field is in the hands of the private sector.
[48:53.060 --> 48:55.900]  So, in this case, the platforms themselves.
[48:55.900 --> 49:01.760]  And there are many projects to try and offer alternatives to mainstream online pornography.
[49:02.000 --> 49:06.920]  Embedding values like feminism, non-conformity of gender, sex workers empowerment, and more.
[49:07.360 --> 49:10.200]  These are all interesting projects.
[49:10.640 --> 49:11.720]  Still niche.
[49:12.060 --> 49:15.140]  Still smaller than the mainstream platforms.
[49:16.240 --> 49:24.760]  And the problem is that the public, or at least for what we have seen, is that the public opinion does not seem to see this issue as a priority.
[49:25.020 --> 49:26.660]  Also to this cultural taboo.
[49:27.060 --> 49:38.080]  So, again, as I said before, we should always remember that we are not talking about an emergency in the sense of a priority that we have to address.
[49:38.080 --> 49:43.660]  We can also decide altogether, as a society or individually speaking, just to ignore it.
[49:44.060 --> 49:49.100]  But we have seen all the monsters that are coming out of this problem if we don't tackle it.
[49:49.320 --> 49:58.380]  We have seen what happens when we don't care about the safety of those who are exposed in these videos.
[49:58.380 --> 50:10.160]  If you look at all the hate that there is online against, well, women, most of the times, when they try to express themselves, to show their body.
[50:10.280 --> 50:14.820]  Or also to all the non-heteronormative orientations.
[50:15.000 --> 50:20.800]  Whenever they show themselves online, there is hatred, there is tension.
[50:20.800 --> 50:39.420]  And this could also be part of the consequences of this huge communication machine, these algorithms that we try to build as good as we can, apparently, that still only reiterate the same models over and over again in an incredibly loud way.
[50:39.680 --> 50:47.140]  Some extremists would even call it, I don't know, means of dissemination of rape culture, but not me.
[50:49.280 --> 50:50.920]  Another solution could be advocacy.
[50:51.460 --> 50:57.000]  Advocacy is, of course, a way to go to institutions and ask for them to solve the problems.
[50:57.280 --> 51:00.600]  And this topic is often not welcomed by these political institutions.
[51:01.200 --> 51:16.800]  Pushing for openness on these issues may escalate in a moral panic, as I said, because on one end you have people that are trying to climb this mountain of the cultural taboo, and on the other end you have people that want to push them and actually stop it altogether from the start.
[51:17.280 --> 51:24.820]  And a blind fight against obscenity and other forms of extremism generally is a toxic debate.
[51:24.820 --> 51:36.360]  It's always very difficult to actually talk about these things, because sex has always been an incredibly powerful core element of our cultures, of our societies, of our history.
[51:36.600 --> 51:38.740]  And so it's not gonna be easy.
[51:40.420 --> 51:45.720]  Oh, and sorry, I forgot to mention, you see that there is, yeah, up left, reactive and proactive.
[51:46.040 --> 51:52.460]  The policy and advocacy solution can be both reactive, so you use them after there is a problem, or proactive.
[51:52.460 --> 51:55.840]  You can actually decide to try and prevent problems from happening.
[51:57.320 --> 52:00.540]  And then there's litigation, which is, of course, only reactive.
[52:00.540 --> 52:06.100]  When somebody violates a norm, you can try and, you know, make them accountable for that.
[52:06.260 --> 52:16.660]  And it seems to be the most reachable approach at the present time is the low-hanging fruit, at least for our experience, allows to bring together law and technology specialists.
[52:17.040 --> 52:25.800]  So, yeah, legal experts and IT specialists can have very interesting conversations about many things, apparently, with this.
[52:26.520 --> 52:35.880]  It involves judges, data protection authorities, and so somehow closes the gap between people and institutions also in this.
[52:35.980 --> 52:38.060]  And sets very important judicial precedence.
[52:39.000 --> 52:42.400]  But what about you, if we still have time to talk about it?
[52:42.780 --> 52:44.180]  Are you aware of any issue?
[52:44.380 --> 52:45.860]  What would you like to investigate?
[52:46.420 --> 52:46.800]  How?
[52:46.960 --> 52:48.440]  Which methodology would you use?
[52:48.520 --> 52:52.140]  If you see any problem in the first place, but I don't know.
[52:53.340 --> 53:02.200]  If you do, and you would like to discuss it with us, we'll be more than happy to know what you think.
[53:02.780 --> 53:08.440]  You can, again, find a lot of material from our project on our website.
[53:08.820 --> 53:12.220]  Tracking.exposed is the main website.
[53:12.520 --> 53:15.120]  For the Pornhub project, it's Pornhub.tracking.exposed.
[53:15.120 --> 53:18.780]  You can find even the slides we just used today.
[53:19.200 --> 53:31.000]  And an academic paper that has been recently published about the analysis of the algorithm that we conducted from using the tools that Julia just presented to you.
[53:31.440 --> 53:33.120]  And that was it.
[53:33.420 --> 53:34.280]  Thank you very much.
[53:34.680 --> 53:36.060]  And I hope you had a good time.
[53:36.960 --> 53:38.640]  Julia, Alex, thank you so much.
[53:40.160 --> 53:42.280]  They'll be available outside if you have questions.
[53:42.280 --> 53:42.460]  Thank you very much.