0:00:00.000,0:00:15.005 34c3 intro 0:00:15.005,0:00:21.070 Herald: All right, it's my great pleasure[br]to introduce to you Mustafa Al-Bassam. 0:00:21.090,0:00:26.500 He's gonna talk about uncovering British[br]spies' web of sockpuppet social media 0:00:26.500,0:00:31.720 personas. Mustafa is a PhD student at the[br]University College in London, studying 0:00:31.730,0:00:37.329 information security and focusing on[br]decentralized systems. Mustafa was a co- 0:00:37.329,0:00:43.921 founder of LulzSec, an hacker activist[br]group some of you might have heard of, and 0:00:43.921,0:00:48.339 with that, please give a warm applause to[br]Mustafa. 0:00:48.339,0:00:55.469 applause 0:00:55.469,0:00:57.920 Mustafa Al-Bassam: Hey. So it seems that 0:00:57.920,0:01:02.489 over the past year we've had a lot in the[br]media about this kind of idea that the 0:01:02.489,0:01:06.070 people that you interact with on Twitter[br]and Facebook and other kinds of social 0:01:06.070,0:01:11.580 media are not necessarily who they say[br]they are, and sometimes not even be, they 0:01:11.590,0:01:16.329 might not even be people at all. They[br]might be bots. And we've heard about how 0:01:16.329,0:01:21.009 this might be used to manipulate people[br]into believing certain things or certain 0:01:21.009,0:01:26.189 ideas. And this has become quite a big[br]topic recently, especially after the U.S. 0:01:26.189,0:01:32.159 presidential elections in 2016, where[br]according to one study, up to one in five 0:01:32.159,0:01:36.030 election related tweets weren't actually[br]from real people. And apparently it's 0:01:36.030,0:01:40.759 it's such a big problem that even the[br]president is being manipulated by, to say, 0:01:40.759,0:01:46.250 bots. But, this has been a kind of[br]activity that has been going on for a very 0:01:46.250,0:01:49.119 long time, and not just from Russia or[br]China. 0:01:49.119,0:01:53.869 The West also engages in these kind of[br]activities including the UK and the US, 0:01:53.869,0:02:00.799 but in other kinds, in other regions. So,[br]today I'm talking about what Britain does 0:02:00.799,0:02:08.038 in this regard. So, in the UK we have a[br]NSA-equivalent intelligence agency called 0:02:08.038,0:02:13.280 GCHQ or Government Communications[br]Headquarters. And their job is basically 0:02:13.280,0:02:20.500 like the UK's version of the NSA: to[br]collect as much information as possible 0:02:20.500,0:02:26.080 through wiretaps and mass surveillance[br]systems. But they also have a subgroup or 0:02:26.080,0:02:31.360 subteam within GCHQ called the Joint[br]Threat Research Intelligence Group or 0:02:31.360,0:02:36.420 JTRIG for short. And what these guys[br]basically do is, its basically a fancy 0:02:36.420,0:02:40.970 name for sitting on Twitter and Facebook[br]all day and trolling online. What they do is 0:02:40.970,0:02:44.860 they conduct what they call Human[br]Intelligence, which is kind of like the 0:02:44.860,0:02:49.840 act of interacting with humans online to[br]try to make something happen in the real 0:02:49.840,0:02:54.390 world. And in their own words one of their[br]missions is to use "dirty tricks" to 0:02:54.390,0:03:00.150 "destroy, deny, degrade and disrupt[br]enemies" by "discrediting" them. And we've 0:03:00.150,0:03:05.400 seen JTRIG has been involved in various[br]campaigns and operations, including 0:03:05.400,0:03:10.090 targeting hacktivist groups like Anonymous[br]and LulzSec, and also protests in the 0:03:10.090,0:03:14.510 Middle East, during the Arab Spring and[br]also the Iranian protest in 2009. 0:03:14.510,0:03:20.620 So, a bit of context to what led me to[br]uncover this stuff and to actually 0:03:20.620,0:03:24.930 research this stuff. So in 2011, I was[br]involved with the with the hacktivist 0:03:24.930,0:03:29.510 group LulzSec. And to refresh your memory,[br]LulzSec was a group that existed during 0:03:29.510,0:03:34.650 the summer of 2011 and hacked into a bunch[br]of US corporate and government 0:03:34.650,0:03:40.211 organizations, like the US Senate, their[br]affiliates and Sony and Fox. And in the 0:03:40.211,0:03:46.180 same year I was arrested, and a year later[br]I was officially indicted on a court 0:03:46.180,0:03:50.680 indictment. But the thing that struck me[br]about this indictment was that there was 0:03:50.680,0:03:55.130 absolutely no mention in this court[br]document about how they managed to 0:03:55.130,0:04:01.130 deanonymize me and my co-defendants. Or[br]how they managed to actually link our 0:04:01.130,0:04:06.820 online identities with offline identities.[br]And I thought it was suspicious because 0:04:06.820,0:04:15.010 our US counterparts, actually, their court[br]indictments had a very lengthy sections on 0:04:15.010,0:04:20.540 how they were caught. For example, when[br]the FBI arrested Jeremy Hammond, his court 0:04:20.540,0:04:25.150 indictment had a, had very detailed[br]information about how those guys social 0:04:25.150,0:04:28.540 engineered him and managed to track him[br]through his IP address and through Tor and 0:04:28.540,0:04:33.600 whatnot. But then, fast forward a year[br]later, Edward Snowden started leaking 0:04:33.600,0:04:39.470 documents about the NSA and GCHQ, and then[br]in 2014, one of those documents or some of 0:04:39.470,0:04:45.600 those documents were released on NBC that[br]showed that GCHQ was targeting hacktivist 0:04:45.600,0:04:49.850 groups like Anonymous and LulzSec. And[br]that makes the a lot of sense in my head. 0:04:49.850,0:04:55.820 Because if GCHQ was involved in this[br]denanonymization process, then they 0:04:55.820,0:04:59.410 wouldn't want to have that in the court[br]indictment, because it would reveal the 0:04:59.410,0:05:03.830 operational techniques.[br]And this is one of the leaked slides from 0:05:03.830,0:05:09.870 GCHQ talking about some of the activist[br]groups they target. One of the people 0:05:09.870,0:05:17.460 they targeted was someone who went by the[br]nickname of "p0ke", who was chatting in an 0:05:17.460,0:05:25.220 IRC channel, a public chat network. And[br]this was a public chatting channel where 0:05:25.220,0:05:30.520 people from Anonymous and other kinds of[br]hacktivists kind of sit and chat about 0:05:30.520,0:05:38.580 various topics and also plan operations.[br]And this person "p0ke" was chatting on 0:05:38.580,0:05:47.490 this channel and boasted that they had a[br]list of 700 FBI agents' emails and phone 0:05:47.490,0:05:55.050 numbers and names. And then it turned out[br]that a GCHQ agent was covertly in this 0:05:55.050,0:06:00.950 channel observing what people were saying.[br]And then the GCHQ agent initiated a 0:06:00.950,0:06:05.510 private message with this person to kind[br]of get more information and to try to 0:06:05.510,0:06:12.210 build a relationship with this person. And[br]the agent asked them what was the site and 0:06:12.210,0:06:16.490 then they just gave that information up[br]and they even gave them a sample of some 0:06:16.490,0:06:22.560 of the leaked information. So it turns out[br]that actually GCHQ was active in these IRC 0:06:22.560,0:06:30.930 networks and chat networks for months if[br]not years and they were in up to several 0:06:30.930,0:06:35.590 hundred channels at a time. They were just[br]sitting there idling. They weren't really 0:06:35.590,0:06:41.450 saying much or actually participating in[br]conversation, except that every few months 0:06:41.450,0:06:46.270 you might notice them say "hey" or "lol"[br]in the chat even though it might be out of 0:06:46.270,0:06:49.360 context of the conversation that was going[br]on, presumably so that they wouldn't get 0:06:49.360,0:06:53.520 kicked off the network because some[br]networks kick you off if you're idling 0:06:53.520,0:06:58.419 there for too long. And then often what[br]they would do is they would private 0:06:58.419,0:07:03.139 message people in rooms to try and[br]corroborate information about activities 0:07:03.139,0:07:07.139 that were going on and being discussed or[br]trying to entrap people by getting them to 0:07:07.139,0:07:13.260 admit to things as we saw with p0ke.[br]And he seemed to be quite a common theme 0:07:13.260,0:07:19.470 that these undercover feds and agents were[br]sitting in these chat rooms. In the 0:07:19.470,0:07:26.389 Europol meeting 2011, where 15 European[br]countries were discussing what they were 0:07:26.389,0:07:31.710 doing to tackle Anonymous and LulzSec,[br]apparently there were certainly undercover 0:07:31.710,0:07:36.520 cops in these channels that had an issue[br]with undercover cops investigating each 0:07:36.520,0:07:40.990 other.[br]laughter 0:07:40.990,0:07:53.280 So the GCHQ agent that was targeting p0ke[br]sent them a link to a BBC news article 0:07:53.280,0:08:01.870 about hacktivists. And, according to this[br]leaked slide, this link enabled GCHQ to 0:08:01.870,0:08:08.610 conduct signal intelligence to discover[br]p0ke's real name, Facebook and email 0:08:08.610,0:08:14.530 accounts etc. It doesn't say exactly how[br]they did that, but it's not that hard if 0:08:14.530,0:08:20.830 they have your IP address on user agent.[br]Back then, in 2011, most websites weren't 0:08:20.830,0:08:25.490 using HTTPS, including Facebook, so if[br]they look up your IP address in XKeyscore 0:08:25.490,0:08:29.520 or the dragnet surveillance system, they[br]can easily see what other traffic is 0:08:29.520,0:08:35.010 originating from that IP address, and what[br]Facebook accounts are connected to that IP 0:08:35.010,0:08:41.948 address for example. But in this in this[br]slide leaked by NBC the URL was redacted, 0:08:41.948,0:08:46.399 but it wasn't very hard to actually find[br]that URL, because these were public 0:08:46.399,0:08:51.029 channels that GCHQ agents were talking in,[br]and people haven't been targeted in 0:08:51.029,0:08:56.470 themselves including myself. We were able[br]to find out what that URL shortener was 0:08:56.470,0:09:01.589 I mean what that website was but[br]which turned out to be a URL shortener so 0:09:01.589,0:09:09.949 the website that was sent to p0ke to click[br]was "lurl.me" and according to 0:09:09.949,0:09:16.950 archive.org, here is a snapshot from[br]"lurl.me" in 2013, just before it went 0:09:16.950,0:09:21.279 offline, that basically showed it was a[br]URL shortening service, it looks like a 0:09:21.279,0:09:28.170 generic URL shortening service. One things[br]I noticed is, the domain name sounds 0:09:28.170,0:09:32.820 like "lure me" which is basically what[br]they were doing, 0:09:32.820,0:09:41.119 because JTRIG had this internal wiki[br]where they listed all the tech tools and 0:09:41.119,0:09:47.149 techniques that they use in the operations[br]and one of the categories that they have 0:09:47.149,0:09:54.999 is "shaping and honey pots" and in that[br]category they have a tool code named 0:09:54.999,0:09:59.200 Deadpool which is described as a URL[br]shortening service and that's what 0:09:59.200,0:10:07.970 "lurl.me" was. We first saw "lurl.me" in[br]2009 - the domain name was registered in 0:10:07.970,0:10:16.040 2009 - and almost immediately it was it[br]was linked tweets about Iranian protests, 0:10:16.040,0:10:21.679 and then it went offline in 2013, shortly[br]after (every sudden) leaks in November, 0:10:21.679,0:10:26.089 but interesting if you look up all of the[br]instances of this URL shortener being used 0:10:26.089,0:10:30.209 in social media and Twitter there's[br]probably about 100-200 instances of it 0:10:30.209,0:10:36.040 being used and every single one of those[br]instances where it was used it was 0:10:36.040,0:10:42.829 associated with political activities late[br]in the Middle East or Africa usually to 0:10:42.829,0:10:49.270 protests. And the majority of the most[br]common were coming from the default 0:10:49.270,0:10:54.220 Twitter accounts with no avatar, with very[br]few tweets and they're accounts that were 0:10:54.220,0:10:59.689 active for only a few months between 2009[br]and 2013. 0:10:59.689,0:11:05.589 One of the techniques, or some of the[br]techniques that JTRIG used, in their own 0:11:05.589,0:11:09.680 words to conduct their operations is[br]includes uploading YouTube videos 0:11:09.680,0:11:13.720 containing persuasive messaging,[br]establishing online aliases with Facebook 0:11:13.720,0:11:18.970 and Twitter accounts, blogs on foreign[br]memberships for conducting human 0:11:18.970,0:11:23.129 intelligence, or encouraging discussion on[br]specific issues, sending spoof emails and 0:11:23.129,0:11:28.189 text messages as well as providing spoof[br]online resources, and setting up spoof 0:11:28.189,0:11:34.850 trace sites and this is exactly what we're[br]going to see in the next few slides and in 0:11:34.850,0:11:39.749 most examples that they use for the[br]operations is they actually targeted the 0:11:39.749,0:11:44.950 entire general population of Iran which is[br]a pretty big target audience of 80 million 0:11:44.950,0:11:48.279 people. According to them,[br]they had several goals in Iran: 0:11:48.279,0:11:53.389 the first goal was to discredit the[br]Iranian leadership and its nuclear program 0:11:53.389,0:11:57.469 Second goal was to delay and disrupt on-[br]line access to materials used in the 0:11:57.469,0:12:00.059 nuclear program. Third Goal was [br]conducting online Human 0:12:00.079,0:12:02.739 Intelligence and the fourth goal was the most 0:12:02.739,0:12:07.589 interesting goal my opinion: Counter[br]censorship. It might seem might sound great 0:12:07.589,0:12:12.769 it might sound like almost like GCHQ is[br]kind of aligned with the motives of the 0:12:12.769,0:12:16.480 Internet freedom community by helping[br]these Iranian activists to evade 0:12:16.480,0:12:18.929 censorship.[br]But we're gonna see it's not really the 0:12:18.929,0:12:24.550 case. The main kind of Iran the main kind[br]of sock puppet accounts on Twitter that 0:12:24.550,0:12:32.009 JTRIG was running during this campaign in[br]2009 was called "2000 Iran 0:12:32.009,0:12:36.519 2009 Iran free".[br]This was the most kind of active Twitter 0:12:36.519,0:12:41.679 account that it had and it had 216 tweets[br]and they also had I kind of like a bunch 0:12:41.679,0:12:46.499 of other accounts that were less active[br]that had default avatars probably just to 0:12:46.499,0:12:51.389 kind of, kind of build up their social[br]network that mostly retweeted things, 0:12:51.389,0:12:57.509 retweeted the same things as a display[br]account but slightly rewarded or even with 0:12:57.509,0:13:00.050 them.[br]And what this Twitter account essentially 0:13:00.050,0:13:07.449 did was in quick succession, over a period[br]of like one or two weeks tweeted a bunch 0:13:07.449,0:13:12.920 of links from this URL shortener for[br]various purposes for to various articles 0:13:12.920,0:13:20.319 on blogs online and they also had actually[br]a blogspot website with like one article 0:13:20.319,0:13:28.709 to kind of expand their network I guess.[br]One of the activities that 2009 Iran free 0:13:28.709,0:13:35.730 and the other sock puppets were doing[br]was they were kind of trying to spread the 0:13:35.730,0:13:42.269 same IP addresses as proxies to Iranians[br]to use as a counter cencorship. So for 0:13:42.269,0:13:48.389 example you can see that they have a list[br]of IP addresses here that will hash like 0:13:48.389,0:13:52.269 Iran election that they can use for[br]protests and they and they might sometimes 0:13:52.269,0:14:01.899 feed links to that to to this proxy is[br]using that URL shortener and this is, this 0:14:01.899,0:14:07.329 is quite concerning because well one of[br]the tools used by JTRIG is also called 0:14:07.329,0:14:12.639 codenamed Molten Magma which is basically[br]HTTP proxy to with the ability to log all 0:14:12.639,0:14:16.910 traffic and perform HTTPS man-in-the-[br]middle because, again, they were they were 0:14:16.910,0:14:20.429 spreading exactly the same IP address all[br]of these all these sock puppet accounts 0:14:20.429,0:14:26.009 were spreading exactly the same IP[br]addresses and same links to Iranians to 0:14:26.009,0:14:33.119 help them to or to allegedly help them to[br]a evade common cencorship. And they were 0:14:33.119,0:14:37.569 even claiming that these for the same[br]proxies used by the Iranian government to 0:14:37.569,0:14:41.249 get around their own firewalls so if they,[br]apparently if they block these proxies 0:14:41.249,0:14:45.619 they will block their own access to the[br]outside world. 0:14:45.619,0:14:50.519 And this is essentially what they are[br]doing here. In this kind of context GCHQ 0:14:50.519,0:14:54.610 is kind of acting like the big bad wolf[br]from Red Riding Hood. We might seem like 0:14:54.610,0:15:02.319 they're helping me but they're also[br]causing you harm in the process. 0:15:02.319,0:15:06.629 And this is a, this is a list that[br]contains a list of some of the techniques 0:15:06.629,0:15:13.319 that JTRIG used. This was also a leaked[br]document and this essentially kills two 0:15:13.319,0:15:18.360 birds in one stone because what they do is[br]at the bottom it says one techniques is 0:15:18.360,0:15:22.370 hosting targets' online communications for[br]collecting signal intelligence as we saw 0:15:22.370,0:15:27.120 with p0ke and which is why they tweet[br]these links using URL shortener so they 0:15:27.120,0:15:32.429 can conduct signal intelligence on people[br]who are interested in clicking these 0:15:32.429,0:15:38.839 things and also providing online access[br]uncensored materials and sending instant 0:15:38.839,0:15:42.759 messages to specific individuals giving[br]them instructions for accessing uncensored 0:15:42.759,0:15:47.120 websites.[br]One of the forums that these proxies were 0:15:47.120,0:15:53.939 posted in was whyweprotest.net and someone[br]actually kind of almost got it right. 0:15:53.939,0:15:56.779 Someone asked: 'Why does the government use[br]proxies? That doesn't make any sense, they 0:15:56.779,0:15:59.509 wouldn't need any proxies." And then[br]someone replied: "The Iranian government 0:15:59.509,0:16:03.999 allegedly has set up proxies to monitor[br]connections with from within Iran to be 0:16:03.999,0:16:08.100 able to pinpoint the people who are trying[br]to bypass these blocks." So they're almost 0:16:08.100,0:16:10.569 right because it wasn't the Iranian[br]government that was actually monitoring 0:16:10.569,0:16:18.760 connections in Iran. It was GCHQ.[br]There were also set up, I agree, basic 0:16:18.760,0:16:25.529 websites, that basically acted as RSS[br]feeds to English websites about Iran to 0:16:25.529,0:16:29.629 presumably, but also for counter[br]censorship reasons. One of the same 0:16:29.629,0:16:34.889 things they did was mimic government[br]officials. So for example they might 0:16:34.889,0:16:39.980 post in a forum saying: "Attention users[br]outside Iran, you can call the president 0:16:39.980,0:16:43.839 at this number to discuss the elections[br]direct." And they were hesitant that you 0:16:43.839,0:16:49.829 should not call this number if you are in[br]Iran. And then they will also give an 0:16:49.829,0:16:55.670 email address for the vice president on[br]the Twitter. 0:16:55.670,0:17:00.370 This also matches up with another[br]technique that JTRIG uses, again according 0:17:00.370,0:17:06.549 to the leaked documents, where they send[br]spoof emails and text messages from a fake 0:17:06.549,0:17:11.669 person or mimicking a real person to[br]discredit, promote, distrust, dissuade, 0:17:11.669,0:17:16.829 deceive, deter, delay or disrupt. Whatever[br]the purpose was, they certainly managed to 0:17:16.829,0:17:20.810 promote distrust because one of the[br]replies to this post was: "This can't be 0:17:20.810,0:17:24.599 the president's number because if it were[br]the second call would be answered by 0:17:24.599,0:17:29.850 Iranian intelligence services. So these are[br]strange days. I suppose anything could 0:17:29.850,0:17:33.760 happen at this point."[br]So that was most of the activity that we 0:17:33.760,0:17:40.450 saw in 2009. There was a bunch of other[br]Twitter accounts with default egg, default 0:17:40.450,0:17:46.461 avatars associated with these links. You[br]can find them if you search lurl.me with 0:17:46.461,0:17:52.570 quotation marks and Google with sites[br]-twitter.com. In 2010 there was absolutely 0:17:52.570,0:18:00.120 no activity on Twitter or all social media[br]associated with this URL shorter. Then, in 0:18:00.120,0:18:08.750 2011, we saw some activity in Syria for[br]this URL shortener for a similar purpose 0:18:08.750,0:18:12.620 of conducting censorship resistance in[br]Syria. And they were essentially doing the 0:18:12.620,0:18:18.100 same thing, same techniques, giving people[br]IP addresses to connect to, that you 0:18:18.100,0:18:24.019 thought that they probably are MITM'd.[br]But one of the things they did here as 0:18:24.019,0:18:28.270 well was they didn't just tweet stuff they[br]also posted a YouTube video, like a very 0:18:28.270,0:18:33.150 poorly made YouTube video with only[br]300 views to try to get people to watch 0:18:33.150,0:18:37.600 that. They didn't really try very hard[br]here because if you actually look at the 0:18:37.600,0:18:43.340 times on when these accounts tweeted,[br]all the accounts in Syria actually should 0:18:43.340,0:18:49.750 have tweeted. The only tweet between 9 to[br]5 p.m. UK time Monday to Friday. 0:18:49.750,0:19:00.070 laughter, applause[br]I mean, I think, I don't know I think 0:19:00.070,0:19:06.269 they were lazy, or they were just, they[br]didn't really bother or weren't motivated. 0:19:06.269,0:19:10.700 But one of the limitations that JTRIG has,[br]they actually had one in the leaked 0:19:10.700,0:19:15.549 documents, that they had was they had a[br]list of limitations that the staff have 0:19:15.549,0:19:19.470 when conducting its operations. And one of[br]them is that they have difficulty in 0:19:19.470,0:19:24.549 maintaining more than a small number of[br]unique multi-dimension active aliases 0:19:24.549,0:19:29.880 especially with doing online human[br]intelligence. Which is why we only see 0:19:29.880,0:19:35.130 like one main twitter account for these[br]events and then like a bunch of other kind 0:19:35.130,0:19:38.610 of default expat accounts, usually like[br]five or six. We didn't tend to see 0:19:38.610,0:19:44.460 hundreds of them you only see about less[br]than 10, because this was back in 2009, 0:19:44.460,0:19:50.270 2011. They weren't doing it in an[br]automated way. And they also said the lack 0:19:50.270,0:19:55.559 of continuity in maintaining an alias or[br]communicating via an alias if a staff 0:19:55.559,0:20:02.350 member is away and his or her work is[br]covered by others and also the other one 0:20:02.350,0:20:08.620 was lack of photographs, visual images, of[br]aliases which is why we always see like 0:20:08.620,0:20:12.280 egg or default avatars for these[br]sock puppet accounts because they can't 0:20:12.280,0:20:16.630 unless they have like a full fledge[br]graphics team or have faces of people to 0:20:16.630,0:20:22.120 put in there and they can't really put[br]anything as avatar. They also apparently 0:20:22.120,0:20:28.220 had a lack of sufficient number and varied[br]cultural language advisors eg in Russian, 0:20:28.220,0:20:32.090 Arabic and Pashto which is why we see[br]here on these Twitter accounts they're 0:20:32.090,0:20:36.299 basically tweeting the same thing over and[br]over again with no variation. Here's the 0:20:36.299,0:20:40.249 same text over and over again because they[br]don't have lots of translators to 0:20:40.249,0:20:48.390 translate that.[br]The other thing we saw in 2011 was a very 0:20:48.390,0:20:54.179 targeted attack during the Bahrain[br]protests. They had a twitter account 0:20:54.179,0:21:00.490 called 'Freedom4Bahrain' and this, it just[br]sent two tweets, mentioning two accounts 0:21:00.490,0:21:07.050 "14FebTV" and "14FebRevolution", and[br]these were two accounts that were, 0:21:07.050,0:21:09.470 like,[br]really big kind of social media outlets in 0:21:09.470,0:21:15.460 Bahrain that were covering the protests[br]that were going on there. And these were 0:21:15.460,0:21:21.770 targeted mentions of the kind that we saw[br]with P0ke, so, presumably also here, they 0:21:21.770,0:21:23.809 were using that to conduct Signal[br]Intelligence, 0:21:23.809,0:21:32.019 to discover who was running these two[br]accounts. In 2012 you also saw no activity 0:21:32.019,0:21:42.009 associated with that URL shortener. During 2013 I managed[br]to find one tweet related to Kenya, to the 0:21:42.009,0:21:47.340 Kenyan imposed national politics and this[br]person isn't an education sock puppet, this 0:21:47.340,0:21:52.700 person is a research assistant at the[br]Human Rights Watch. So this, but that begs 0:21:52.700,0:21:58.080 the question of how did he actually get[br]this URL? Probably a similar message to 0:21:58.080,0:22:02.720 P0ke, they probably sent him a link[br]through a private message found that 0:22:02.720,0:22:08.460 interesting and tweeted it, so not only[br]are they targeting protesters, they are 0:22:08.460,0:22:16.750 also targeting NGOs. Then, in 2013,[br]all of the infrastructure associated with 0:22:16.750,0:22:23.370 URL-shortener was shot offline, this was[br]in 2013, which was a few months after the 0:22:23.370,0:22:26.790 Edward Snowden leaks, so they had a bit of[br]delay of doing it, but it must have been a 0:22:26.790,0:22:32.840 real pain in the arse for them to have to[br]renew all their infrastructure, but I did 0:22:32.840,0:22:38.340 do some digging into some of other host[br]names that were hosted on this lurl.me 0:22:38.340,0:22:44.820 server. Between 2009 and 2013, most of[br]these host names seem to be random 0:22:44.820,0:22:51.090 alphanumeric, the main names, and some of[br]them are using publicly the DNS providers 0:22:51.090,0:22:57.350 like DynDNS or DNSAlias, I wasn't able to[br]find any websites archived for these 0:22:57.350,0:23:02.039 domains, so it doesn't seem that there was[br]any websites there, but if you have any 0:23:02.039,0:23:06.250 ideas let me know, because one of the[br]things that I suspect is that these might 0:23:06.250,0:23:09.809 have been malware endpoints or command[br]control servers, that they were using, so 0:23:09.809,0:23:13.880 if you have any and monitoring tools or[br]logs then maybe you should look up some of 0:23:13.880,0:23:18.759 these host names. But one of the[br]interesting domain names that I thought 0:23:18.759,0:23:25.049 was interesting there was dunes[br]adventures.net and this is the archived 0:23:25.049,0:23:27.009 page for Dunesadventures 0:23:27.009,0:23:29.440 which was another[br]website based in Kenya. They were up to 0:23:29.440,0:23:35.110 something in Kenya and it claimed that[br]they were having this was a very basic one 0:23:35.110,0:23:41.009 page website that was kind of very poorly[br]made and they claimed that they were 0:23:41.009,0:23:44.539 having site problems and apparently "we[br]have noticed problems with our booking 0:23:44.539,0:23:49.220 system, this has been taken offline until[br]our techs find the problem - we apologize 0:23:49.220,0:23:53.250 for any inconvenience". but there was never[br]any booking system in the first place, 0:23:53.250,0:23:58.270 this was just pretty much a ruse to make[br]it look like if you go to this website, a 0:23:58.270,0:24:03.360 legitimate company was hosting there. So[br]if you mind anything about that, then I'd 0:24:03.360,0:24:08.139 be curious as well. I also if there's any[br]GCHQ agents in the room and then I'm 0:24:08.139,0:24:15.779 happy to get drink with you as well.[br]That's all I have for today, does anyone 0:24:15.779,0:24:26.960 have any questions?[br]applause 0:24:26.960,0:24:41.510 (Herald) asks for questions[br](Mic Question): OK, IRC asks: Deceiving 0:24:41.510,0:24:46.350 a target into trusting you and leaking any form[br]of infos is used everywhere right now, IRC, 0:24:46.350,0:24:50.970 Twitter and Facebook and so on. How would you[br]advise people to distinguish between a 0:24:50.970,0:24:54.059 genuine identity and an undercover agent? 0:24:54.059,0:24:56.029 (Speaker): "I think that's a very good[br]question because- 0:24:56.029,0:24:59.121 (H.): So just just a quick second, if you 0:24:59.121,0:25:03.400 really have to leave the room right now,[br]people, please do so quietly, we still 0:25:03.400,0:25:08.019 have a talk going on and it's really[br]unrespectful if you make that much noise 0:25:08.019,0:25:13.190 and interrupt this whole thing.[br]applause 0:25:13.190,0:25:17.300 I know a lot of people are interested in[br]the talk afterwards but we'll all get you 0:25:17.300,0:25:18.300 in and sorry. 0:25:18.300,0:25:23.309 (S.): So I think I was very good question[br]because if you're conducting, if you're 0:25:23.309,0:25:26.990 doing activism online and you need to be[br]anonymous and you dont want to meet up 0:25:26.990,0:25:30.450 with people in person, then how do you[br]know that the people you communicating 0:25:30.450,0:25:34.350 with, or if you are like in a public group[br]where you personally accept new members 0:25:34.350,0:25:39.490 into that group, how can you put, how do[br]you know or kind of differentiate between 0:25:39.490,0:25:44.299 who's actually there to harm your group or[br]who's actually there to contribute? I 0:25:44.299,0:25:51.250 think the answer there lies in, what you[br]share. Don't share information that comes 0:25:51.250,0:25:55.690 with anyone that could potentially put you[br]at harm, even with people that you trust, 0:25:55.690,0:25:59.409 so essentially don't trust anyone and[br]this is a basic OP Sec rule. This is 0:25:59.409,0:26:06.799 how Jeremy Hammond messed up a few years[br]ago, because they caught him, because he 0:26:06.799,0:26:11.259 was revealing too much information about[br]his life, like where where he eats or 0:26:11.259,0:26:18.759 something like that or his previous drug[br]records and they were able to use that to 0:26:18.759,0:26:22.940 kind of figure out who he was and that was[br]the same mistake that P0ke made he, was 0:26:22.940,0:26:30.299 too open and friendly to that agent for no[br]reason. So I think the kind of answer is 0:26:30.299,0:26:34.590 to do your operations in a way where you[br]dont have to trust people. 0:26:34.590,0:26:40.409 (Mic Question): "How effective do you 0:26:40.409,0:26:45.350 think these methods are, because we've[br]seen the number of followers on Twitter 0:26:45.350,0:26:50.350 and the number of views on YouTube were[br]very low so, how much people can, is 0:26:50.350,0:26:51.970 affected by this kind of operations" 0:26:51.970,0:26:57.730 (S.): Yes, so there was also a slide I[br]meant to put in there, that was leaked page 0:26:57.730,0:27:03.110 another leaked page from GCHQ that had a[br]list of bullet points on what they 0:27:03.110,0:27:07.370 considered to be an effective operation[br]and some of those bullet points include 0:27:07.370,0:27:11.929 how many people click that link, how many[br]people, how many people watch the youtube 0:27:11.929,0:27:15.120 video, etc, so it's pretty much the same[br]ways that you would measure it how many 0:27:15.120,0:27:19.889 people viewed a specific message. Now in[br]their specific use cases I don't think 0:27:19.889,0:27:23.820 they were very successful on a large[br]scale, specifically in Iran protests 0:27:23.820,0:27:27.499 because the Twitter accounts had very few[br]followers and their YouTube videos only 0:27:27.499,0:27:33.279 had a few hundred views but they might[br]have been, obviously more succesful in 0:27:33.279,0:27:37.039 more target cases when targeting specific[br]individuals by doing the Bahrain case or 0:27:37.039,0:27:38.039 the p0ke case. 0:27:38.039,0:27:39.610 (H.): over there please. 0:27:39.610,0:27:45.220 (Mic Question): Sure, thank you, so I'm[br]just curious if you were familiar with the 0:27:45.220,0:27:49.730 work of Erin Gallagher, she's done work to[br]try to figure out, kind of quantitatively 0:27:49.730,0:27:52.809 and make these visualizations, to try to[br]figure out if a particular twitter account 0:27:52.809,0:27:57.279 for example is a bot or whether it's a[br]person and there's some you know rules of 0:27:57.279,0:28:00.499 thumb regarding like, you know if the bots[br]just kind of interact with each other and 0:28:00.499,0:28:01.909 don't react, don't interact with real[br]people 0:28:01.909,0:28:07.340 im just curious what, what techniques you[br]may know of to, to figure out you know 0:28:07.340,0:28:10.539 what is a bot and what is not and whether[br]you are familiar with those particular 0:28:10.539,0:28:11.559 lines of a research. 0:28:11.559,0:28:16.960 (S.): I'm not familiar with with their[br]work, but thank you all check out. In terms 0:28:16.960,0:28:24.140 of what kind of metrics that you could use[br]or to use to see if a account is valid or 0:28:24.140,0:28:29.720 not, I mean, I think, I guess they're,[br]their tweeting kind of, habits and when 0:28:29.720,0:28:34.010 they tweet for example could be[br]indicative, so for example we saw this 0:28:34.010,0:28:38.251 person only tweet at 9 to 5. Obviously[br]that's quite easy to make that it's on the 0:28:38.251,0:28:44.120 case and also I think one useful things[br]might be might be interesting to do, is 0:28:44.120,0:28:50.879 try to map the network of these accounts.[br]If you like build up like a web of 0:28:50.879,0:28:55.909 followers, that you might be able to very[br]easy for graphically detect, very obvious 0:28:55.909,0:28:59.100 clusters for accounts that are following[br]each other, to be to be very signal. 0:28:59.100,0:29:01.370 (Mic): Yeah for sure, thank you. 0:29:01.370,0:29:04.440 (H.) Lets switch over to mic 6 please 0:29:04.460,0:29:05.309 (Mic 6 question): Thank you for the- 0:29:05.309,0:29:11.580 thank you for the great talk, how would[br]you compare the former British activities 0:29:11.580,0:29:18.149 to the current Russian activities, maybe a[br]talk in itself, but... 0:29:18.149,0:29:20.429 (S.) To be honest, I haven't been digging 0:29:20.429,0:29:23.919 too deep in the details or following too[br]much about the Russian activities, so I 0:29:23.919,0:29:26.860 can't really comment about that, I don't[br]know how prolific it is, I only mentioned 0:29:26.860,0:29:31.760 it briefly in the beginning of the slides[br]because it was to give some context, so 0:29:31.760,0:29:34.370 I'll have to research more to the Russian[br]activities. 0:29:34.370,0:29:39.020 (H.) Go to mic 5 again 0:29:39.020,0:29:42.140 (Mic 5 Question): Thanks, to continue 0:29:42.140,0:29:51.830 from the person who spoke, that would have[br]been my question. So, just to add up onto 0:29:51.830,0:29:58.860 that, did you stumble upon similar[br]patterns coming from say Canberra or a 0:29:58.860,0:30:00.230 Washington DC? 0:30:00.230,0:30:05.440 (S.): So these accounts were very[br]specific to just to the UK expressions, 0:30:05.440,0:30:09.280 there was no kind of collaboration there[br]with other countries within the five eyes, 0:30:09.280,0:30:15.200 like the US or Australia, but I think they[br]might have, 0:30:15.200,0:30:19.120 GCHQ I think has collaborated with the NSA 0:30:19.120,0:30:23.060 JTRIG specifically I think has collaborated[br]before with the NSA to delegitimize 0:30:23.060,0:30:27.929 certain people. So for example[br]we saw during a few years ago or last year 0:30:27.929,0:30:34.230 I think there was a drone attack, someone[br]was illegally killed in a drone strike in 0:30:34.230,0:30:40.220 Iraq, he was a suspected to be an ISIS[br]member, Junaid Hussain, and apparently the 0:30:40.220,0:30:45.299 way that he was deanonymized or the way they[br]found this location is that the US, the 0:30:45.299,0:30:49.269 FBI specifically, had an informant that was[br]talking to this person and that informant 0:30:49.269,0:30:53.480 sent them and sent them a link that was[br]generated by GCHQ and then since that link 0:30:53.480,0:30:56.710 they were able to deanonymize them so I[br]think there's some collaboration there but 0:30:56.710,0:30:59.110 this is mostly UK activity. 0:30:59.110,0:31:04.315 (H.): Last question, we are out of time.[br]Thank you again, Mustafa. applause 0:31:04.315,0:31:31.940 subtitles created by c3subtitles.de[br]in the year 2019. Join, and help us!