[00:01.950 --> 00:02.810]  Hello, everyone.
[00:03.230 --> 00:08.870]  As you just heard, this is Privacy Badger and Panopticlick versus the trackers round one.
[00:09.990 --> 00:20.210]  So, as a quick introduction to... whoops, click on here... and then...
[00:24.370 --> 00:27.130]  As soon as we figure out how slideshows work, we'll give her a talk.
[00:27.350 --> 00:30.210]  I use a different slideshow software than Cooper does, so...
[00:30.210 --> 00:30.610]  There you go.
[00:30.610 --> 00:31.090]  Okay, cool.
[00:31.490 --> 00:34.850]  As an introduction to me, my name is Bill Budington.
[00:35.330 --> 00:42.150]  I work at the EFF on various projects, one of which is HTTPS Everywhere, which you might know.
[00:42.590 --> 00:45.770]  Another of which is Panopticlick, which we'll talk about in a minute.
[00:47.870 --> 00:49.070]  I'm Cooper Quinton.
[00:49.270 --> 00:51.330]  I'm also a technologist at EFF.
[00:51.630 --> 00:52.970]  I work on Privacy Badger.
[00:53.130 --> 00:57.550]  I'm running a CTF for EFF here at HOPE, which you should check out if you haven't.
[00:57.850 --> 00:59.870]  And I do security research as well.
[01:02.590 --> 01:03.550]  So... wait.
[01:03.730 --> 01:04.510]  Oh, that's me.
[01:06.710 --> 01:07.890]  So, what is EFF?
[01:08.410 --> 01:11.290]  How many people here have not heard of EFF?
[01:13.150 --> 01:13.690]  All right.
[01:13.810 --> 01:14.990]  Somebody has not heard of EFF.
[01:15.150 --> 01:15.410]  Awesome.
[01:15.890 --> 01:20.010]  And that's good, because the next part of this talk won't work, otherwise, if everybody has heard of EFF.
[01:20.970 --> 01:33.650]  So, earlier this year, my colleague, Jeremy Galula, did some tests which proved that T-Mobile was throttling traffic for all video streams if the customer had binge-on service enabled.
[01:34.230 --> 01:40.750]  We got on Twitter, we published a blog post, and then we got on Twitter to ask the T-Mobile CEO, John Legere, about this.
[01:41.030 --> 01:43.930]  And he had this lovely response for us.
[01:44.570 --> 01:49.250]  I think that these questions are actually important, so I want to take a minute to respond to him.
[01:52.730 --> 01:54.810]  So, who the f*ck is EFF, anyway?
[01:56.610 --> 02:04.390]  Well, he could have looked at our Twitter profile, he could have Googled it, he could have asked a hacker, but that's okay.
[02:04.590 --> 02:05.470]  I can explain it to him.
[02:05.910 --> 02:08.630]  We're a nonprofit which has been around for 25 years.
[02:08.910 --> 02:13.510]  Our mission is to make sure that when you get online, your rights come with you.
[02:13.510 --> 02:15.790]  And we do this in a few different ways.
[02:17.090 --> 02:27.570]  We have an amazing legal team, which has been involved in many legal cases against the National Security Agency for their unconstitutional spying on Americans and civilians all over the world.
[02:28.310 --> 02:32.190]  The legal team... Yeah, round of applause for them, they're amazing.
[02:34.590 --> 02:41.950]  The legal team's involved in dozens of cases each year, which range from filing amicus briefs to testifying in front of the Supreme Court.
[02:43.890 --> 02:55.290]  Also, members of the legal team run our Coders Rights Project, which focuses on keeping security researchers and hackers from getting into trouble due to their research and defending them when they do.
[02:56.270 --> 03:01.690]  I should take a note here to also mention that we are not lawyers, and none of this constitutes legal advice.
[03:01.870 --> 03:03.950]  If you need legal advice, go find one of our lawyers.
[03:07.050 --> 03:09.130]  So, why are we stirring up so much trouble?
[03:09.430 --> 03:11.250]  Well, one, because it's our job.
[03:11.370 --> 03:13.070]  It's what our members expect us to do.
[03:13.950 --> 03:19.230]  And secondly, because there's a lot of broken legislation, a lot of bad tracking technology out there.
[03:19.230 --> 03:22.490]  And if we didn't do this, we'd be remiss.
[03:23.190 --> 03:29.490]  Our members really count on us to help the Internet, keep the Internet safe and secure.
[03:31.950 --> 03:42.790]  So, we do things with the activism team at EFF, like fly blimps over the NSA, pointing downwards, saying illegal spying happening here.
[03:43.930 --> 03:44.910]  That was great.
[03:45.190 --> 03:46.330]  Parker was on that blimp.
[03:46.590 --> 03:48.370]  You can talk to him at the EFF booth.
[03:50.150 --> 03:53.910]  The blimp was loaned to us by Greenpeace, so we gotta give them credit too, for sure.
[03:54.530 --> 03:54.610]  Yeah.
[03:55.530 --> 04:15.610]  And we have campaigns against things like the Computer Fraud and Abuse Act, which you might know of, which, you know, basically categorizes all sorts of activity as criminal hacking and makes it very easy to prosecute hackers, very vulnerable hackers on any number of accounts.
[04:15.870 --> 04:24.210]  This was enacted, to note, in 1986, after a Hollywood craze of, you know, around the movie War Games.
[04:24.550 --> 04:32.210]  So, this kind of shows you how Hollywood and Congress kind of line up when they're prosecuting vulnerable populations.
[04:36.070 --> 04:51.710]  We also have an international team, which works on things like the Vassanar Agreement, which is an international multilateral treaty, which makes possession of hacking tools illegal.
[04:52.070 --> 04:56.250]  So, there's a lot of great international work that we do as well.
[05:00.620 --> 05:03.400]  So, John's final question, who pays us?
[05:03.820 --> 05:13.420]  It's a good question, and it turns out that the bulk of our budget comes from our over 25,000 amazing members, some of whom have been with us for over 25 years.
[05:14.200 --> 05:17.640]  Our members took to Twitter to inform John Legere of this fact.
[05:19.180 --> 05:24.280]  And let him know that if they had to choose between T-Mobile and EFF, they would happily choose EFF.
[05:28.240 --> 05:31.000]  After all this, John Legere quickly backpedaled.
[05:31.500 --> 05:34.120]  Apparently, he does know who EFF is after all.
[05:34.840 --> 05:35.800]  So, hey, that's great.
[05:39.960 --> 05:45.840]  We also have a technology projects team, which both Cooper and I are a part of, which makes the Internet more secure.
[05:46.220 --> 05:47.740]  We do this in two ways.
[05:47.960 --> 05:50.880]  We do it from the client side and from the server side.
[05:51.000 --> 06:10.560]  For the clients, we get end users to install things like HTTPS Everywhere, which, as I said, I work on, an add-on, which makes sure that it's an add-on available for Chrome and Firefox, which makes sure that if there is a secure endpoint on the web, that you're actually accessing that secure endpoint for a website that you're going to.
[06:12.720 --> 06:18.320]  And on the server side, we work on things like Certbot, which is formally called Let's Encrypt.
[06:19.520 --> 06:29.440]  And this makes it really easy and free and fast and automated to get certificates that are recognized by browsers on a site that you may run.
[06:31.480 --> 06:35.380]  So, enough with the intro to the EFF, let's talk about surveillance.
[06:36.900 --> 06:42.380]  Before you, you see the schematic for the Panopticon, which is where Panopticlick derives its name.
[06:42.380 --> 06:46.340]  This was first conceptualized by Jeremy Bentham in 1787.
[06:47.180 --> 06:55.020]  And it's a way to kind of keep tabs on all prisoners and watching their every movements from this kind of central node here in the middle.
[06:56.520 --> 07:08.180]  And we thought of this as a great analogy for how online trackers try to really view your every movement as you're traveling and browsing through the web.
[07:10.260 --> 07:17.160]  So, the story of how online trackers started begins in the 90s and really hits the spotlight by 1999.
[07:18.060 --> 07:25.200]  This New York Times article written by Glenn Fleischman of that year laments the fact that browser cookies can be added...
[07:25.200 --> 07:30.960]  can be used as a way to track users as they browse the web through page views.
[07:31.840 --> 07:49.800]  So, the old model of the web as a simple lookup table from URLs to their contents was kind of becoming a thing of past, where instead of this kind of unidirectional server-to-client model, a lot of more information by clients was being sucked up by the servers as well.
[07:49.920 --> 07:58.400]  Of course, IP addresses were already sucked up, but, like, a lot of different information about browsers and browsing habits was, you know, starting to take hold here.
[08:02.120 --> 08:08.620]  So, cookies let us have what's called third-party tracking, and there's a graph above here.
[08:09.140 --> 08:24.660]  The way third-party tracking works is that when you visit a site, say, for example, New York Times, you're not only visiting the New York Times, many other domains also get loaded along with the content from NYT.
[08:24.660 --> 08:33.460]  So, for example, you have advertising, you have analytics, you have things that are...the sole purpose is just to track you.
[08:35.000 --> 08:37.040]  And all of these get loaded on New York Times.
[08:37.180 --> 08:42.040]  They all get to set a cookie in your browser, and they all get to see that you're visiting the New York Times.
[08:43.180 --> 08:52.460]  Then, when you visit another site, like CNN, a bunch more third-parties load, and they all get to set cookies, and they all get to see that you're on CNN.
[08:53.040 --> 09:03.340]  Some of these third parties are loaded on New York Times, and the third parties are represented by triangles here...are loaded on New York Times, CNN, and every other site you visit.
[09:03.680 --> 09:14.580]  These third parties get to put together a picture of all of the sites that you've been to, because they keep getting to read your cookie, And they keep getting to see what site you're getting that cookie on.
[09:15.660 --> 09:28.860]  So, this lets people track you around the web and breaks this sort of first-party origin of the web where you think that you're having this one-to-one conversation, but you're actually having a one-to-one conversation with many people listening in.
[09:34.320 --> 09:37.440]  So, why are we focused on third-party trackers in this talk?
[09:38.600 --> 09:40.340]  Well, there's a few things.
[09:40.600 --> 09:42.040]  One is that they're non-consensual.
[09:42.040 --> 09:48.760]  Like I said, when you go to New York Times, you're not expecting to talk to DoubleClick, Scorecard Research, and a bunch of other people.
[09:48.900 --> 09:50.700]  You're expecting to talk to the New York Times.
[09:51.660 --> 09:52.900]  Also, they're ubiquitous.
[09:53.180 --> 09:55.800]  Almost every website has third-party trackers.
[09:55.940 --> 10:01.240]  And there was a study that showed that almost 90% of news websites had third-party trackers.
[10:02.420 --> 10:03.560]  They're hard to avoid.
[10:03.740 --> 10:05.640]  Most people don't even know that this exists.
[10:05.640 --> 10:13.500]  Or if you do notice it, it's because you shopped for something on Amazon, and then ads for that thing followed you around for the next month.
[10:14.840 --> 10:16.240]  They're hard to avoid.
[10:18.120 --> 10:20.860]  It's not always intuitive how these things work.
[10:21.060 --> 10:24.700]  And installing things like ad blockers doesn't always do the trick.
[10:25.480 --> 10:28.300]  And if there's a strong financial incentive.
[10:29.180 --> 10:32.220]  Third-party tracking is big business.
[10:32.680 --> 10:35.240]  It's a multi-billion dollar industry.
[10:35.760 --> 10:39.060]  And it's been around for a few years, and it's only growing.
[10:42.920 --> 10:45.460]  So, who do we see doing this third-party tracking?
[10:45.660 --> 10:46.580]  There's some key players.
[10:47.140 --> 10:52.240]  Some that you've heard of, like Facebook, and Google, DoubleClick.
[10:53.420 --> 10:55.240]  You've probably heard of Add This, right?
[10:56.260 --> 10:58.600]  Google and DoubleClick, they're doing it for ads.
[10:58.780 --> 11:02.040]  Facebook is doing it for ads, and who knows.
[11:04.680 --> 11:07.920]  Add This is, yeah, doing it for money.
[11:07.920 --> 11:10.760]  And there's invisible ones, like Scorecard Research.
[11:11.000 --> 11:13.700]  Scorecard Research is on a bunch of websites.
[11:14.140 --> 11:17.140]  And I still can't figure out what service they provide.
[11:19.520 --> 11:21.740]  AxiCom isn't on websites, but they're a data broker.
[11:21.900 --> 11:29.100]  They buy data from all these guys, and then sell it back to other people about your browsing habits, what you're buying, et cetera, et cetera.
[11:30.040 --> 11:33.160]  So, third-party tracking is also useful for spies.
[11:34.360 --> 11:44.760]  From the Snowden documents, we've seen that the NSA really likes piggybacking off of Google and Yahoo tracking cookies to help them track people around the web.
[11:46.640 --> 11:48.680]  So, it's a pretty big problem.
[11:51.420 --> 12:01.040]  So, you could see that, you know, previously we had this model of, you know, certain sites having cookies on, you know, just like, you know, tracking them.
[12:01.040 --> 12:03.960]  And, you know, that's, you know, great.
[12:04.060 --> 12:05.660]  Maybe you can actually just clear your browser.
[12:05.900 --> 12:08.580]  So, you know, maybe that's sufficient.
[12:09.560 --> 12:14.060]  So, the model, you know, before, you know, all the cookies all the time.
[12:14.280 --> 12:14.960]  Just gather them all.
[12:15.100 --> 12:16.060]  Just like Cookie Monster, right?
[12:17.740 --> 12:39.600]  But in 2009, ten years later, from the original article I showed, the situation had gotten so bad and the methods so advanced that this article by Ed Felton in Freedom to Tinker nostalgically reminisces on the good old days when trackers would just use cookies to track you instead of all the other things that they're doing now.
[12:40.640 --> 12:45.580]  The sentiment that if you're going to track users, at least be transparent about it.
[12:46.040 --> 12:53.620]  The view of cookies being this transparent way to track users shows how bad things had gotten just by this point.
[12:55.420 --> 12:56.660]  So, what happened?
[12:56.900 --> 13:00.480]  How did things get this bad over the span of a single decade?
[13:00.880 --> 13:11.720]  Well, new techniques were being developed that utilize the pervasive inclusion of browser plug-ins such as Flash, Java, and Silverlight as persistent stores of data.
[13:12.560 --> 13:16.060]  And they do this in a very underhanded and covert manner.
[13:16.940 --> 13:32.020]  In addition to being new stores of persistent data, which are called super cookies, by the way, they're also very difficult to purge because they circumvent the normal mechanisms built into the browser for purging your cookies.
[13:32.820 --> 13:40.400]  So, you'll have to, for instance, in Flash, you have to go to, you know, Flash settings, internal Flash, and delete them that way.
[13:40.400 --> 13:56.760]  And so there's all these different plug-ins that have much more fine access and much, you know, much more privilege than the browser would ordinarily give add-ons, say, which is the distinction of add-on and plug-in.
[13:58.820 --> 14:07.520]  What's worse, these plug-ins can collude with one another in respawning cookies if they're not deleted simultaneously on every single one of them.
[14:07.520 --> 14:13.160]  In 2010, security researcher, Sami Kamkar, came up with this idea of the ever cookie.
[14:13.820 --> 14:33.820]  And what the ever cookie is, is say you delete cookies on Flash and Silverlight, but not in Java, then there's this Java code that uses and immediately repropagates the cookies to the other two platforms, and maybe even your browser cookies on your browser itself.
[14:34.560 --> 14:44.640]  In this way, these super cookies turn into these very, you know, persistent ever cookies, and they achieve a level of persistence that's very, very difficult to break.
[14:47.300 --> 15:02.620]  In addition to actual data stores, trackers started to rely on these little bits of information left by your browser every time you access a site, such as fonts, and also the headers that you send upon every web request.
[15:03.720 --> 15:12.020]  You're able to, they're able to basically combine these bits of information into a cohesive whole or this unique fingerprint of your browser.
[15:13.360 --> 15:26.120]  These, so basically if you use both cookies and fingerprinting, then trackers can really, you know, have this kind of pervasive view of your actual presence on the web.
[15:26.120 --> 15:29.820]  It can actually, you know, follow you in a very fine-grained way.
[15:31.760 --> 15:36.800]  So we were confident at EFF that browser fingerprinting was possible.
[15:37.340 --> 15:44.460]  And as data-driven technologists, we wanted to gather more concrete statistics and learn more about the information left by users' browsers.
[15:44.460 --> 15:52.760]  So in January 2010, we asked volunteers to participate in this experiment that we branded Panopticlick.
[15:53.520 --> 15:59.300]  The site, when you visited it, did these fingerprinting techniques, you know, in an opt-in way.
[15:59.300 --> 16:06.660]  And users were able to determine what exactly their fingerprint looked like.
[16:06.860 --> 16:14.300]  And it also allowed us to gather statistics about web trends in general and how fingerprintable users' browsers are.
[16:15.820 --> 16:18.320]  So a quick math detour here.
[16:19.280 --> 16:21.100]  I might lose you, but I'll bring it back.
[16:21.920 --> 16:27.580]  So a uniqueness property, when we talk about uniqueness, that's measured in the form of entropy.
[16:28.360 --> 16:31.580]  Entropy is a quantity measured in bits, so it's log base 2.
[16:32.020 --> 16:39.440]  In order to know how many bits you need to uniquely identify someone, you take the log base 2 from the population that you're measuring from.
[16:40.520 --> 16:48.840]  So, for instance, to determine how many bits of entropy you need to find out the identity of someone on Earth.
[16:49.500 --> 16:58.080]  Well, you need 32.7 bits of entropy because 2 to the 32.7 is 7.1 billion, or about how many people there actually are on Earth.
[17:00.460 --> 17:17.900]  So the change in entropy is this quantity, or this term that's called surprisal, because it's basically measuring how surprising a new bit of information is when you learn something new about someone, how much does that level of entropy change over time,
[17:18.120 --> 17:20.060]  or when you learn something new about someone.
[17:20.290 --> 17:24.460]  So that's called surprisal, and it's determined from this equation here.
[17:25.860 --> 17:28.640]  Now, this is kind of best illustrated by example.
[17:30.820 --> 17:47.220]  The surprisal, or the amount that the entropy changes, if you learn that someone is born on, or is a Capricorn, is log base 2 of the proportion of the population, that's a PR, parentheses, that actually is a Capricorn.
[17:47.220 --> 17:51.020]  Or about log base 2 of 1.12, or about 3.58 bits.
[17:53.080 --> 18:04.220]  The, you know, change in entropy when you learn someone is born on January 2nd, is log base 2 of 1 out of 365, or about 8.51 bits.
[18:04.220 --> 18:16.980]  So, you can chain together independent facts about someone, to really kind of, to add up these bits of entropy, and get a good idea of how, of what your exact identity is.
[18:17.500 --> 18:19.940]  These need to be independent facts about someone.
[18:20.440 --> 18:22.760]  In this case, these are not independent facts.
[18:22.900 --> 18:25.740]  If you know someone's born on January 2nd, you already know that they're Capricorn.
[18:25.740 --> 18:29.960]  So, it's not adding anything new to your knowledge about that person.
[18:32.530 --> 18:39.720]  So, with Panopticlick, the independent facts that we measured were a combination of headers, and JavaScript-detected properties.
[18:40.400 --> 18:45.080]  And the population size being measured was everyone that ever took the test.
[18:46.060 --> 18:48.240]  The pool of all volunteers, basically.
[18:49.320 --> 18:53.820]  Users that had JavaScript turned off were obviously better protected in this study.
[18:53.820 --> 19:00.540]  In that case, we only had to rely on the header information that was being delivered upon the web request.
[19:04.630 --> 19:07.090]  In May 2010, we published our findings.
[19:07.970 --> 19:13.710]  In this paper, by our chief technologist, Peter Eckersley.
[19:13.990 --> 19:21.150]  And we described how the overwhelming majority of users' browsers had this uniquely identifiable fingerprint.
[19:21.970 --> 19:26.590]  In fact, 84% of people's browsers had uniquely identifiable fingerprints.
[19:27.950 --> 19:34.210]  Along with... if you had Flash installed on your browser, this number actually jumped to 94%.
[19:34.210 --> 19:50.210]  So, this was kind of a good way to drive home the point that users really needed to take concrete steps in order to protect their browsers and their ability to feel safe against trackers in their browsing.
[19:52.270 --> 20:02.270]  If this wasn't bad enough, in the intervening years since we originally published the Panopticlick findings, we've seen more and more advanced forms of tracking appear.
[20:02.630 --> 20:05.490]  Show of hands, how many people know what this illustration means?
[20:05.910 --> 20:06.290]  Yeah?
[20:06.470 --> 20:06.830]  A few of you?
[20:06.850 --> 20:07.090]  I do, I do.
[20:07.090 --> 20:11.630]  So, this is something called canvas fingerprinting.
[20:12.370 --> 20:17.890]  And it basically renders text onto an HTML5 canvas element.
[20:18.670 --> 20:32.950]  And even little changes in your fonts or your operating system configuration, things like anti-aliasing or kerning or whatever, result in different images being rendered here.
[20:34.310 --> 20:43.810]  And if you take that image and serialize it and run a hashing function over it, then that's a pretty good metric of how unique your browser is.
[20:44.550 --> 20:53.870]  Using a study with Mechanical Turk, they were able to determine that 5.7 bits of entropy are determined by this alone.
[20:54.870 --> 21:01.210]  And this was first... this is, I think, that first premiered on a paper called Pixel Perfect in 2012.
[21:01.850 --> 21:08.570]  And since then, even by 2014, it was widely, widely implemented by trackers.
[21:11.010 --> 21:18.970]  You can see here that this is an example of canvas fingerprinting by a library, open-source library called fingerprint or fingerprint 2.
[21:20.350 --> 21:33.110]  So, in order to maximize the chances of the generated image being unique, you can see this kind of complex overlay of different shapes, colors, fonts, UTF characters.
[21:33.110 --> 21:40.190]  And this, you know, increases the chances of you having a unique canvas fingerprint based on this image.
[21:41.130 --> 21:45.630]  Over time, canvas fingerprinting has really gotten only smarter and smarter.
[21:46.510 --> 21:53.610]  And it became a big problem when trackers AddThis and Ligitus implemented it in 2014.
[21:54.330 --> 22:00.570]  As much as 5% of the Alexa top 100,000 sites were found to be tracking users in this manner.
[22:01.330 --> 22:11.250]  A white paper released in July 2014 titled, The Web Never Forgets, really shown a spotlight on these kind of tracking techniques.
[22:11.790 --> 22:20.910]  And after a flurry of media attention, these top two canvas fingerprinters actually stopped doing it and stopped doing the shady practice.
[22:24.250 --> 22:25.710]  And so, a quick shout-out.
[22:25.850 --> 22:32.630]  One of the authors of that paper, Ganesh Akhar, actually came and worked on Privacy Badger and contributed a patch to detect canvas fingerprinting.
[22:33.850 --> 22:36.590]  So, this is web audio fingerprinting.
[22:36.730 --> 22:48.090]  And it's pretty similar to canvas fingerprinting in that it uses the HTML5 web audio API to create a sound wave and then read that back and serialize it.
[22:48.090 --> 22:55.970]  Due to differences in your hardware, due to differences in your sound card, etc., it will get a pretty unique fingerprint.
[22:59.790 --> 23:02.450]  And then there's online to offline tracking.
[23:02.770 --> 23:07.390]  And this is sort of the latest direction, the newest trend for the ad tech industry.
[23:07.530 --> 23:15.810]  The demand to link devices to specific real-world identities and link online and offline shopping and viewing habits.
[23:15.810 --> 23:19.390]  So, this company called Silver Push took a crack at this.
[23:19.770 --> 23:27.250]  They took a frequency which was inaudible to human ears and encoded it in television commercials.
[23:27.730 --> 23:34.690]  The frequency, even though it was inaudible to you, it was not inaudible to your phone or probably your dog.
[23:36.850 --> 23:38.270]  They're tracking my dog?
[23:40.550 --> 23:41.110]  Soon.
[23:43.290 --> 23:50.630]  So, they included a library in mobile applications which would detect these signals coming from your TV.
[23:50.870 --> 23:52.950]  I know, this sounds like I should put on a tinfoil hat.
[23:53.690 --> 23:59.270]  And call back to the Silver Push servers letting them know that the application had heard this signal.
[23:59.270 --> 24:04.650]  So, therefore, you are watching this commercial and here's all this other data from your phone about who you are.
[24:05.710 --> 24:07.970]  Apparently, this was a tracker too far.
[24:08.350 --> 24:13.550]  And the FTC investigated them and got them to stop using this technology.
[24:13.870 --> 24:17.970]  But they might not be the only ones using this technology or similar technologies.
[24:19.330 --> 24:24.150]  Facebook also dabbled in this because it turns out Facebook knows a lot about who you are.
[24:25.710 --> 24:37.750]  So, they developed this program called Atlas which would let them link the ads that you had seen online to purchases that you had made offline.
[24:38.270 --> 24:56.870]  When you use things like loyalty cards or store accounts associated with your phone number, when you use your credit card if that's linked to Facebook, they would take that data and link that to the ads you had seen on Facebook, letting them know if an ad had caused you to go buy something offline.
[24:57.230 --> 25:03.750]  This, as far as I know, is still active, although I think that I've heard it hasn't been very successful for them as a program.
[25:04.310 --> 25:06.770]  But maybe they're just planning to iterate it.
[25:06.870 --> 25:07.550]  We're not really sure.
[25:11.690 --> 25:21.190]  So, having heard about all the ways that your browsing habits and devices can be tracked might leave you with a serious case of privacy nihilism.
[25:21.570 --> 25:26.610]  You might be feeling right now that privacy is hopeless and why should you even bother?
[25:27.830 --> 25:30.050]  And I think it's important, actually.
[25:30.230 --> 25:35.330]  I want to take... I think we need to take a minute here to address some of the common forms of privacy nihilism.
[25:37.510 --> 25:42.330]  So, one thing I often hear is, but I like targeted ads.
[25:42.610 --> 25:43.130]  They're great.
[25:43.250 --> 25:44.690]  They help me find things I want.
[25:45.770 --> 25:46.330]  Really?
[25:46.330 --> 25:47.970]  They don't help me like that, but okay.
[25:49.110 --> 25:56.170]  The problem is, you have no control over how your information is stored or used.
[25:57.170 --> 26:01.430]  These third parties have no obligation to delete the data that they collect.
[26:01.430 --> 26:04.970]  They have no obligation to temporarily store the data they collect.
[26:05.270 --> 26:09.350]  And they certainly don't have any obligation to make sure that the data is correct.
[26:11.570 --> 26:13.530]  The data can also be stolen.
[26:13.870 --> 26:14.810]  It can be sold.
[26:14.990 --> 26:19.070]  Even if you think you trust the company, you might not trust the company in the future.
[26:19.990 --> 26:21.830]  And it can be misused.
[26:22.690 --> 26:43.150]  A couple of months ago, a woman named Sharona Coates wrote a story for Rewire about this guy who was selling the ability to geo-target advertisements to women at abortion centers, trying to convince them to not get abortions.
[26:44.110 --> 26:46.450]  This is hugely invasive.
[26:47.030 --> 26:56.450]  And he was able to do this through the standard geo-targeting abilities of Facebook and other ad service providers, right?
[26:56.570 --> 27:04.710]  Using mobile location and then seeing ads as you're browsing Facebook or Tumblr or whatever at the Planned Parenthood.
[27:04.710 --> 27:09.250]  I mean, this is so problematic in so many ways.
[27:09.930 --> 27:15.790]  And we're just seeing the tip of the iceberg of this sort of thing.
[27:15.970 --> 27:20.970]  I think that there's a lot more terrible things that could be done with this.
[27:21.530 --> 27:25.150]  So, I don't like targeted ads.
[27:29.210 --> 27:39.270]  So, some people would like to talk about how privacy is dead, which is really interesting because when you look at who is saying it, it turns out to be this guy.
[27:40.830 --> 27:42.010]  Who you might know.
[27:43.910 --> 27:49.230]  But Mark Zuckerberg, CEO of Facebook, likes to declare privacy is dead.
[27:49.230 --> 28:03.310]  And yet he turns around and buys $30 million worth of houses that surround his house in order to have just the privacy that he says everyone else shouldn't have.
[28:04.250 --> 28:13.210]  So, when he talks about privacy being dead, what he really means is that privacy is dead for normal people, but not for the ultra-rich tech elite.
[28:15.850 --> 28:18.570]  So, why should you actually care about privacy?
[28:19.010 --> 28:20.670]  Well, there's kind of a lot of reasons.
[28:21.190 --> 28:27.770]  Maybe you want to read articles or download books which are controversial in nature or maybe just embarrassing.
[28:27.950 --> 28:29.390]  Maybe they have to do with a medical condition.
[28:30.450 --> 28:43.190]  Maybe you want to protect yourself against little bits of data that when put together could be more embarrassing or lead to more of a cohesive picture of who you are and what your life is.
[28:44.050 --> 28:47.830]  Maybe you want to avoid geo-targeting of ads.
[28:48.050 --> 28:52.290]  For instance, what Cooper just mentioned, the visiting of an abortion clinic.
[28:52.870 --> 29:06.410]  Or maybe you just want to avoid the chilling effects of speech entailed when you know that someone is always looking over your shoulder and gathering data about you.
[29:09.770 --> 29:13.110]  So, privacy lets us make mistakes.
[29:13.370 --> 29:15.590]  It lets us play with ideas.
[29:16.210 --> 29:18.790]  It lets us grow as individuals.
[29:19.290 --> 29:23.790]  Who here has ever had a thought that they don't want shared with the world?
[29:24.170 --> 29:24.690]  Yeah.
[29:25.710 --> 29:29.050]  I mean, everybody's like thinking of their thought now and shamed.
[29:31.770 --> 29:35.610]  Privacy gives us the space to grow and define who we are.
[29:35.610 --> 29:38.870]  And it's hugely important to a free society.
[29:40.450 --> 29:42.270]  So, don't give up hope.
[29:42.510 --> 29:42.870]  Yeah?
[29:43.070 --> 29:43.230]  HOPE.
[29:43.370 --> 29:43.630]  Get it?
[29:43.710 --> 29:44.470]  That's the conference we're at?
[29:46.070 --> 29:47.150]  That was my joke.
[29:47.270 --> 29:47.590]  I'm sorry.
[29:49.210 --> 29:50.670]  He made me tell it.
[29:53.950 --> 30:01.230]  So, we can actually categorize the privacy protective efforts into three distinct categories.
[30:01.230 --> 30:03.750]  The good, the bad, and the ugly.
[30:05.970 --> 30:08.090]  So, for the good, Flash is dying.
[30:08.450 --> 30:09.970]  And it's dying really quickly.
[30:11.090 --> 30:11.650]  Yeah.
[30:11.770 --> 30:12.330]  And good riddance.
[30:15.800 --> 30:22.600]  In the last five years, the appearance of Flash has gone from just below 50% to just above 20%.
[30:22.600 --> 30:25.240]  And it's likely this trend will continue.
[30:25.480 --> 30:26.500]  And it's not just Flash.
[30:26.500 --> 30:35.660]  Across the board, the old plugins that had that system-level access that browsers didn't want to give are going the way of the dinosaurs.
[30:36.760 --> 30:42.980]  So, that means, at least theoretically, the browser APIs can be utilized to rein in some of the worst abuses of the past.
[30:45.760 --> 30:56.880]  Other good efforts to stop tracking include the Tor browser, which makes your browser look like every other instance of the Tor browser that's using it currently.
[30:58.260 --> 31:03.620]  This basically makes tracker's data collection ineffectual completely.
[31:04.940 --> 31:08.460]  Caveats are that it's not the most intuitive thing to use.
[31:10.340 --> 31:15.140]  It's also that, you know, it might be slow for some people.
[31:15.400 --> 31:16.040]  We've heard that.
[31:17.820 --> 31:27.800]  And on the upside, a lot of the privacy protective measures that have been included in the Tor browser are making their way upstream to Firefox.
[31:28.280 --> 31:31.580]  There's an active effort to bring those patches back to Firefox.
[31:31.980 --> 31:34.820]  So, there's good news on the horizon.
[31:36.500 --> 31:40.920]  Also, Firefox tracking protection is a great way.
[31:41.460 --> 31:44.440]  This is on by default within private browsing mode.
[31:45.060 --> 31:54.680]  And this uses the disconnect and tracker blocking mechanisms in order to protect you when you're using private browsing mode.
[31:56.980 --> 32:01.100]  Also, research projects like OpenWPM are great.
[32:01.880 --> 32:10.260]  OpenWPM basically tracks first parties and their inclusion of third party trackers across time and gathers that data.
[32:10.960 --> 32:19.300]  And basically, you know, scrapes the Alexa top X number of sites every month to figure that out and get those web trends.
[32:19.300 --> 32:23.620]  So, more research in this area is really needed.
[32:24.280 --> 32:28.580]  And the more data we have about trackers, the more effectively we can combat them.
[32:31.320 --> 32:38.860]  Then we have efforts that are bad at stocking trackers, like incognito mode in Chrome.
[32:38.860 --> 32:45.620]  To be fair, incognito mode is not really intended to be used against trackers.
[32:45.860 --> 32:51.540]  It's intended to prevent local data about your browsing from being stored on your own computer.
[32:53.340 --> 32:54.700]  Ad blockers...
[32:54.700 --> 32:57.620]  Oh, yeah.
[32:57.740 --> 33:02.680]  And it doesn't really prevent those pesky fingerprinting techniques that we saw earlier.
[33:03.520 --> 33:09.800]  Ad blockers are also a pretty bad method generally at preventing trackers, although there are some exceptions.
[33:10.580 --> 33:19.720]  Since most of them don't even block trackers directly, and they don't block them at all sometimes, especially the invisible ones.
[33:20.440 --> 33:25.940]  Sometimes, ad blockers don't even block ads from being loaded on the page.
[33:26.240 --> 33:30.780]  They simply stop them from being displayed on the page that you're viewing.
[33:31.060 --> 33:36.320]  So, in that case, the ads are still able to track you just as well as they were before.
[33:40.660 --> 33:48.380]  And often, they have questionable business models, such as some that will remain unnamed.
[33:49.960 --> 33:50.280]  Boo!
[33:51.200 --> 33:52.900]  Boo those unnamed guys!
[33:54.100 --> 34:03.920]  And also, the W3C's do not track policy gives users a way to kind of flag that they do not wish to be tracked by two trackers.
[34:03.920 --> 34:10.340]  But there's no enforcement mechanism whatsoever with this policy.
[34:11.200 --> 34:15.000]  So, there's no way for trackers to actually...
[34:15.000 --> 34:18.800]  I mean, they could obey it, but there's no incentive for them to do so.
[34:21.770 --> 34:23.430]  So, then we have the ugly.
[34:23.650 --> 34:26.330]  And these are incentives by the ad industry.
[34:27.030 --> 34:28.950]  And yeah, totally feel free to boo these guys.
[34:30.110 --> 34:34.870]  So, there's the ad choices program by the Digital Advertisers Alliance.
[34:34.870 --> 34:41.490]  And what this does is basically advertisers have proposed to self-regulate.
[34:43.050 --> 34:46.590]  The Digital Ad Alliance offers members an opt-out.
[34:48.190 --> 34:53.650]  But the opt-out isn't actually an opt-out from tracking.
[34:53.910 --> 34:56.810]  It's only an opt-out from seeing targeted advertisements.
[34:56.810 --> 34:59.070]  So, you still get tracked.
[34:59.430 --> 35:01.990]  You just don't see the results of it.
[35:02.650 --> 35:07.730]  There's still no requirements on what data they can or cannot collect in store.
[35:07.990 --> 35:09.610]  It's not legally binding.
[35:10.290 --> 35:14.250]  It doesn't address any of the security concerns like malvertising.
[35:14.250 --> 35:19.110]  And it's still only has limited adoption.
[35:20.210 --> 35:21.410]  The other...
[35:21.410 --> 35:22.350]  Oh, too far.
[35:25.750 --> 35:31.130]  The other advertising initiative is from the Interactive Advertising Bureau.
[35:31.450 --> 35:32.910]  And they have two initiatives.
[35:33.170 --> 35:34.170]  One is called Deal.
[35:34.490 --> 35:36.750]  And this is sort of the ad blocker blocker.
[35:36.830 --> 35:43.210]  You might have gone to Wired or Forbes and seen this pop-up that says, Hey, you're using an ad blocker so you can't read our site.
[35:43.510 --> 35:46.590]  Turn off your ad blocker and then you can come read our site.
[35:47.530 --> 35:49.830]  Again, this doesn't address tracking.
[35:50.010 --> 35:51.910]  It doesn't address people's privacy concerns.
[35:52.510 --> 35:56.490]  And it still doesn't address people's security concerns over things like malvertising.
[35:58.150 --> 36:01.750]  It's annoying for users and it's patronizing, right?
[36:02.110 --> 36:03.930]  Hey, you're using an ad blocker.
[36:04.010 --> 36:05.550]  You didn't know that that hurts us.
[36:05.730 --> 36:07.390]  Yeah, it did, but your ads suck.
[36:12.530 --> 36:14.230]  It's annoying for users.
[36:14.570 --> 36:16.830]  And maybe it's more like deal with it.
[36:17.890 --> 36:21.670]  Lean is the program for less obnoxious ads.
[36:23.150 --> 36:25.710]  So, Lean still doesn't address...
[36:25.710 --> 36:34.850]  The idea behind Lean is the ad industry's attempt that they think, OK, your problem with ads is that they're obnoxious, they're full video, they play in the center of the page.
[36:35.090 --> 36:36.990]  Yeah, sure, that's kind of a problem.
[36:36.990 --> 36:39.990]  But Lean still doesn't address privacy concerns.
[36:40.670 --> 36:45.370]  It only minimally addresses security concerns by requiring that ads be served over HTTPS.
[36:45.990 --> 36:50.830]  So, great, we want to encrypt everything, but that's not really the heart of the concern with ads.
[36:53.600 --> 36:54.860]  Oh, too far.
[36:55.100 --> 36:56.060]  Too far back.
[36:57.720 --> 37:02.180]  So, none of these really addressed the concerns that we had at EFF.
[37:02.180 --> 37:05.200]  So, we did what we always do.
[37:05.400 --> 37:08.320]  We combined technology, law, and activism.
[37:09.060 --> 37:13.840]  And we came up with Privacy Badger, Panopticlick, and Do Not Track.
[37:17.160 --> 37:19.440]  This is Privacy Badger in its natural habitat.
[37:19.900 --> 37:19.920]  Aw.
[37:20.180 --> 37:20.500]  Yeah.
[37:23.280 --> 37:26.860]  So, Privacy Badger is a browser extension.
[37:26.860 --> 37:29.220]  It's free and open-source software.
[37:29.740 --> 37:34.800]  It focuses on completely blocking trackers from even connecting to your browser.
[37:35.880 --> 37:36.680]  Uh, it...
[37:37.440 --> 37:49.520]  Unlike most, uh, tracker blocking and ad blocking technology, which uses a blacklist, Privacy Badger uses a heuristic to try to figure out dynamically what's tracking you in particular.
[37:50.720 --> 37:55.380]  And it lets honest actors, people that aren't actually trying to track you, a way out.
[37:57.500 --> 38:05.020]  Privacy Badger tells sites you do not wish to be tracked by sending the, uh, W3C dnt equals one header.
[38:05.540 --> 38:09.880]  It then looks for third parties that get loaded as you browse the web.
[38:09.880 --> 38:23.440]  Uh, if a third party is seen on several different domains, and it appears to be tracking you, say, by setting high-entropy cookies, fingerprinting you, setting high-entropy supercookies, then it gets blocked.
[38:25.100 --> 38:38.700]  So, this is Privacy Badger running on Gawker.com, and you can see that it's, uh, so all the domains that are in red are being blocked, uh, that's the domains to where the slider's to the left, if you're red-green colorblind.
[38:39.000 --> 38:46.540]  Uh, and the domains that are green are not being blocked yet, because Privacy Badger hasn't seen them tracking across multiple sites.
[38:49.840 --> 38:50.500]  Um...
[38:50.500 --> 38:50.740]  Yeah.
[38:50.920 --> 38:52.900]  For some sites, we...
[38:52.900 --> 38:57.560]  For some third parties, we don't want to block them entirely.
[38:58.480 --> 39:11.220]  Uh, things like the Creative Commons image server, things like Google Maps, things like YouTube embeds, people want these as part of their web browsing experience, but they're still able to potentially track you.
[39:11.440 --> 39:20.600]  So, for these things, we have what we call a cookie block list, where we allow them to load, but we try to prevent as many vectors of tracking as possible.
[39:20.800 --> 39:31.740]  So, we block them from setting and reading cookies, we block them from certain types of supercookies, uh, I mean, as many as we can, and we block them from as many types of fingerprinting as we can.
[39:32.160 --> 39:38.000]  Uh, this works pretty well to let people have the web they want while still not being tracked as much as possible.
[39:39.360 --> 39:42.500]  Another of the core tenets of Privacy Badger is user choice.
[39:42.840 --> 39:48.620]  Uh, if you don't like the decisions that Privacy Badger has made, uh, we want you to be able to override those.
[39:48.900 --> 39:53.260]  So, you can enable or disable Privacy Badger on a given site if it's not working.
[39:53.560 --> 40:01.420]  Uh, and you can choose to block, cookie block, or allow any given domain, uh, if you disagree with Privacy Badger's decision.
[40:03.380 --> 40:23.680]  Privacy Badger also replaces certain social widgets like Facebook button, like buttons, tweet buttons, SoundCloud widgets, with a click to play button, so that the Facebook and Twitter widgets aren't tracking you around the web, but you still get the usefulness of being able to like things and tweet about them.
[40:24.300 --> 40:26.420]  Uh, and this work was done by Franzi Rosner.
[40:28.580 --> 40:39.940]  So, what about third-party sites that legitimately don't wish to track users, but still need to set a cookie for their technology to work, or do some other thing for their technology to work?
[40:40.740 --> 40:45.780]  So, for those, we have our policy side, which is our do not track policy.
[40:45.960 --> 40:48.540]  This is different from the W3C's do not track policy.
[40:48.540 --> 40:56.980]  So, we've written a document, which states that users sending the DNT header won't be tracked.
[40:57.240 --> 40:58.720]  And I'll explain what that means in a second.
[40:59.820 --> 41:05.300]  It's posted, people, anybody can post it at a specific location on their website.
[41:06.120 --> 41:17.160]  Um, and we think that because you're posting this document on your website, if you violate the terms, the FTC can take action against somebody who does this.
[41:21.020 --> 41:25.980]  The other side of this, uh, so what it does is it says user identifiers will be discarded.
[41:26.720 --> 41:30.920]  Uh, logs will not be kept longer than necessary, a specific determined amount of days.
[41:31.200 --> 41:34.280]  I think it's like seven or nine, uh, in the document.
[41:34.680 --> 41:40.420]  Uh, data can be kept for debugging or security until it's no longer necessary, then it must be destroyed.
[41:40.420 --> 41:47.400]  And data can be anonymized or aggregated for analytics into sufficiently large buckets of people.
[41:48.340 --> 41:59.540]  Um, so the other side of this is that sites that adopt the DNT policy are automatically white listed by Privacy Badger.
[41:59.540 --> 42:03.780]  And other participating tracking protection software.
[42:04.820 --> 42:19.140]  Um, blocking sites, what we think, our strategy here, is that blocking sites that don't respect people's privacy, blocking sites that don't respect DNT, creates an incentive for sites to respect DNT.
[42:19.260 --> 42:20.900]  Because they want to be able to show their ads.
[42:22.360 --> 42:24.910]  Or, or, you know, hopefully not ads, but...
[42:25.480 --> 42:28.620]  Um, so, right now we have a policy up here.
[42:28.950 --> 42:34.080]  And it's been adopted by DuckDuckGo, Adzerk, Mixpanel, Medium, Disconnect, and a bunch of other people.
[42:36.500 --> 42:42.260]  So, um, last year, late last year in December, we launched Panopticlick 2.0.
[42:42.840 --> 42:55.260]  Um, with this second iteration of Panopticlick, we focused on bringing a new suite of, uh, tests to show how well your ad and tracker blocking and protection software is actually working.
[42:55.700 --> 43:03.620]  In order to get accurate results, we've set up a number of domains that kind of look like trackers, but are actually just us.
[43:04.260 --> 43:10.260]  Um, and, uh, see how well your, your protection software sizes up against them.
[43:11.020 --> 43:22.240]  Um, the resources in, uh, domains that are included in this way, um, are, are kind of included in this way that tries to trigger three different types of blockers.
[43:22.480 --> 43:27.180]  Those that use domain blacklisting, so, you know, just full domains that are, that are blocked.
[43:27.580 --> 43:34.220]  Um, those that use URL fragments, such as, like, you know, uh, add underscore URL equals in the URL.
[43:34.640 --> 43:37.760]  And also, um, heuristic blockers like privacy badger.
[43:38.200 --> 43:51.680]  Um, if your protection software is triggered and actually blocks the mock trackers that we've set up, um, then we can know that your tracker blocking and ad blocking software is actually working properly.
[43:52.220 --> 44:00.420]  Um, if your protection software is misconfigured, then we know that, um, you need better protection.
[44:00.940 --> 44:08.100]  And, uh, we kind of gently nudge you to either install privacy badger or something else that's appropriate for your platform.
[44:08.100 --> 44:13.560]  Uh, we've indicated also in these test results, and the third one is kind of fuzzy, I know.
[44:13.840 --> 44:18.120]  Um, but that says, um, does your browser accept the do not track policy?
[44:18.280 --> 44:25.040]  So, do they unblock, um, sites and domains that have posted the do not track policy on their, on their site?
[44:25.720 --> 44:39.560]  Um, and finally, we've radically simplified the fingerprinting results from Panopticlick 1.0 to make sure that non-technical users can get a good at-a-glance look at, and see how, uh, unique their browser is.
[44:39.880 --> 44:41.280]  Um, but, don't worry.
[44:41.420 --> 44:48.340]  We have the full fingerprinting results behind a single click, and, um, if you want to be really diligent about it, then you can easily do so.
[44:49.900 --> 44:54.550]  Um, in addition to a tracker and ad blocking tests, um...
[45:00.460 --> 45:00.740]  No?
[45:01.180 --> 45:01.660]  There you go.
[45:02.040 --> 45:02.140]  Wait.
[45:03.880 --> 45:04.620]  Ah, there we go.
[45:05.260 --> 45:09.760]  Uh, we've rewritten the back end completely in Python Flask.
[45:09.760 --> 45:13.540]  So, you can actually see how, what it's doing, and how it's working.
[45:14.300 --> 45:29.660]  We've, you know, all these, all of our, you know, projects at AFF are on GitHub, and you can easily clone them, and you can actually, you know, set up a, really easily set up a, a Panopticlick instance yourself, because we've made it, uh, dockerized,
[45:30.040 --> 45:32.280]  and you can, you know, kind of run Docker in and run it yourself.
[45:32.280 --> 45:37.520]  Um, we've also added six new fingerprinting metrics on our own, Panopticlick.
[45:37.980 --> 45:42.430]  Um, Canvas fingerprinting, OpenGL fingerprinting, which is kind of similar to Canvas fingerprinting.
[45:43.240 --> 45:53.160]  Um, also header-based, um, metrics like DNT, uh, header, language, platform, and browser touch support, um, which is a JavaScript, um, property.
[45:53.600 --> 45:59.360]  Uh, and these kind of give you a more accurate idea of, um, your browser's uniqueness.
[45:59.360 --> 46:04.880]  Uh, so we've, we made the fingerprinting results, uh, sorry.
[46:04.980 --> 46:20.380]  We've also made these fingerprinting results epoch, so we're basically measuring your browser up against browsers that we've seen recently, instead of, like, the browsers that we measured, you know, six years ago when we first started Panopticlick.
[46:20.720 --> 46:27.060]  Um, that's to get a better idea of how unique your browser is right now, rather than everything that we've ever seen.
[46:28.380 --> 46:36.920]  Um, so, since launch, we've seen the tracker test run over 800,000 times, which is a great, great success for us.
[46:37.500 --> 46:50.700]  Um, we've seen, uh, add, uh, and tracker protection, uh, on, you know, improve on 15,500, uh, unique IPv4 addresses, which is awesome.
[46:50.700 --> 47:05.420]  Um, if this count on the bottom looks small compared to the top, we'll consider that if you changed IP addresses between installing the protection software and testing again, um, or if you're using a VPN, it's only counted once, or if you're using Tor,
[47:05.580 --> 47:06.320]  it's only counted once.
[47:06.940 --> 47:09.800]  Um, so there's a lot of things that we can't measure like that.
[47:10.220 --> 47:12.720]  Um, so we've seen a lot of success with this tool.
[47:12.860 --> 47:13.780]  Yay, ponies!
[47:13.920 --> 47:14.060]  Yay!
[47:15.180 --> 47:17.080]  Uh, but we're not done yet.
[47:17.080 --> 47:19.900]  We have some plans for the future in store.
[47:20.320 --> 47:29.120]  Uh, first of all, we're planning to open up the, uh, anonymized data as best we can, because, obviously, this is dealing with people's private data, too.
[47:29.320 --> 47:31.220]  So we want to be very careful about that.
[47:31.760 --> 47:35.960]  Um, so we can take advantage of all the data that we've collected over the years.
[47:36.060 --> 47:47.000]  In addition, we, um, started using it as a testing framework for Privacy Badger, um, to make sure that when Privacy Badger blocks domains, it's actually acting properly and doing it in the right way.
[47:47.860 --> 47:54.560]  Um, and in the future, we hope other browsers will, um, will adopt Panopticlick for their testing.
[47:57.220 --> 48:15.100]  So, uh, in Privacy Badger and DoNotTrackLand, what we would like, uh, to do in the future is improve the heuristic, uh, reduce false positives, detect and block more types of super cookies and fingerprinting, uh, and, uh, to get wider D&T adoption.
[48:16.920 --> 48:18.740]  So, what do we do from here?
[48:18.880 --> 48:20.240]  Oh, sorry, I stole your slide, Bill.
[48:21.540 --> 48:32.760]  So, um, first of all, you can help us by actually using Privacy Badger, using Panopticlick, and if you, uh, have friends or family that you want to be protected, encourage them to use it as well.
[48:33.100 --> 48:42.880]  Um, it's better to get everyone you know using these tools, right, because the better protection they have, the better, the more safe and, and private web we have as, you know, together.
[48:43.600 --> 48:54.880]  Um, also, um, if you, uh, you know, you adopt the D&T policy on your own sites, the more the D&T policy is adopted, the more teeth it has, really.
[48:55.480 --> 48:57.340]  Um, help us out with code.
[48:57.860 --> 48:58.900]  Um, donate to EFF.
[48:59.100 --> 49:03.940]  These are important projects which we would love to continue, and only with your member dollars can we actually do it.
[49:06.940 --> 49:09.540]  So, uh, we need better tools in the browser.
[49:10.060 --> 49:10.240]  Still.
[49:10.500 --> 49:12.980]  Uh, I can't, we can't do everything in extension land.
[49:13.180 --> 49:15.560]  Uh, we need built-in tracking protection.
[49:15.720 --> 49:16.820]  Firefox and Opera have it.
[49:16.900 --> 49:17.360]  That's great.
[49:18.000 --> 49:20.900]  Uh, we need, uh, double keyed cookies and super cookies.
[49:21.060 --> 49:25.840]  This is a cookie which is keyed also to the third party and also to the first party that was seen on.
[49:25.960 --> 49:29.360]  So, the cookie would be keyed to double click on New York Times.
[49:30.020 --> 49:33.720]  Um, we need browsers which are hardened against fingerprinting.
[49:33.900 --> 49:35.780]  Some of this is work is being done in Firefox.
[49:36.120 --> 49:37.780]  A lot of it has been done in the Tor browser.
[49:37.780 --> 49:43.300]  But it needs to be happening in Chrome, it needs to be happening in Safari, and it needs to be happening in Opera.
[49:43.960 --> 49:47.860]  Uh, we also need better controls for blocking and clearing super cookies.
[49:48.460 --> 49:50.900]  Uh, how many of you know how to clear Flash super cookies?
[49:51.100 --> 49:51.980]  I can't see hands.
[49:51.980 --> 49:54.640]  But, yeah, like one or two people probably, right?
[49:54.760 --> 49:55.980]  I don't actually know how to do it.
[49:56.040 --> 49:58.260]  I think you have to install Flash to clear them.
[49:58.520 --> 50:01.260]  You have to go to a Flash applet on Macromedia.com.
[50:03.020 --> 50:03.900]  It's a nightmare.
[50:04.720 --> 50:07.720]  Uh, we also need new business models for the web.
[50:07.880 --> 50:11.440]  The web, until now, has been based on tracking ads.
[50:11.780 --> 50:13.240]  But it doesn't have to be.
[50:13.800 --> 50:19.260]  Uh, it could be, we could try out memberships, donations, uh, crowdfunding.
[50:19.520 --> 50:25.000]  If a potato salad can get crowdfunded for hundreds of thousands of dollars, I think your website can do just fine.
[50:25.840 --> 50:26.800]  Uh, micropayments.
[50:26.880 --> 50:30.100]  There's a lot of really cool work going on there that I would like to see take off.
[50:30.500 --> 50:33.560]  Uh, or at the very least, non-intrusive advertising.
[50:33.780 --> 50:34.940]  Have display advertising.
[50:35.120 --> 50:36.940]  Have ads that don't track people.
[50:38.520 --> 50:43.580]  Uh, so we don't want to leave you with the feeling, uh, leaving this room of despair.
[50:44.060 --> 50:45.700]  Um, that's not our goal.
[50:45.860 --> 50:48.060]  We want to really tackle a hard problem.
[50:48.420 --> 50:51.940]  But, like this quote says, um, the situation isn't hopeless.
[50:52.280 --> 50:53.240]  Um, it's far from it.
[50:53.480 --> 50:55.800]  And we're working on tools to make it much better.
[50:56.780 --> 50:58.760]  Um, so don't be a privacy nihilist.
[50:58.840 --> 50:59.700]  Be a privacy vegan.
[50:59.820 --> 51:04.340]  We need adamant people about, out there to, you know, protect yourself on the web.
[51:04.700 --> 51:06.680]  And, um, you know, help us do it.
[51:07.760 --> 51:10.660]  So, is advertising the best way to fund the web?
[51:11.480 --> 51:12.460]  It's hard to say.
[51:12.960 --> 51:15.320]  Um, it's certainly here right now.
[51:15.760 --> 51:16.840]  Maybe we can change that.
[51:17.080 --> 51:18.480]  Maybe we can't.
[51:19.420 --> 51:27.500]  But if advertising is the best way to fund the web, and if it's what we're going to live with, it must stop violating users' privacy.
[51:28.580 --> 51:29.160]  Thank you.
[51:37.760 --> 51:40.240]  So, I think we have no time for questions.
[51:40.440 --> 51:43.460]  Is there somewhere we could post up to answer questions if people have them?
[51:45.940 --> 51:46.220]  Okay.
[51:46.480 --> 51:48.940]  So, if you have questions for us, uh, come out to the lobby.
[51:49.100 --> 51:49.600]  We'll be out there.
[51:50.000 --> 51:50.500]  Thank you very much.
[51:50.840 --> 51:51.520]  Thank you very much.