Skip to main content

tv   Arthur Holland Michel Eyes in the Sky  CSPAN  August 21, 2019 4:55am-6:27am EDT

4:55 am
left in the drawing board. >> good afternoon. i want to welcome those of you who are here in the auditorium. here in washington dc as well as those of you who are watching online. on our website or on c-span. i believe they are in fact on site here. according to them. just wanted to go over a few
4:56 am
items before we get to the meat of the program here. i would ask that everybody here turn off their cell phones or at least have them in a silent mode and those other electronic items as well. we will have a q&a portion as we get close to the end of this, and when it goes down, it will ask folks to please wait to be called upon and wait for the microphone so everyone in the audience has an opportunity to hear you. that i will ask you to announce your name and affiliation. our topic today is this particular absolutely fascinating and terrifying book eyes in the sky, by arthur hall and michelle and i have to.out that the digital version that when i call the eye of saffron i will compare it to the lord of the rings.
4:57 am
it was literally six years ago this month, that an nsa contract whistleblower by the name of edward snowden, burst upon the world scene with his absolutely amazing revelations. about mass government survey blades and being taken place on the so-called war on terror error. dozens of not hundreds of stories on snowden and his risen revelation that out from the 2013, and continuing to that talk about that specific day. that's all been about electronic surveillance. in terms of the listening variety. our cell phone conversations, our text messages and things of that nature. our guest today, brings us but maybe as scary or even scarier technological news which is the tom cruise minority report
4:58 am
scenario is not exactly so far fetched anymore. in fact the technology we are going to talk about is actually inspired by a different movie which i won't steal his thunder on. let me introduce our guests. over here in the far right is our guest today arthur hall and michelle was a journalist researcher and founder and codirector of the center of the drone bar college and arthur is written for us news, fastcompany, motherboard, the list goes on and on. he is a co-author of the drone primer a of key issues and close encounters in the international airspace. sitting directly next to me as jenna mclaughlin who is a national security and investigation report of our yahoo! news. she focuses on the intelligence community foreign policy and other issues. jenna has previously covered national security for cnn foreign policy the intercept and mojo's following her graduation from john hopkins in 2014. in between, next to josh is
4:59 am
shaun vega whose policy county the demand progress. john has served as legislative counsel for policy and insured for foundation and a google policy fellow for public representation. in addition to sort of policy, the domain progress fund, he also served as the director which he helped covan on capitol hill. you may have heard of a law firm in california. tadpole of texas. an analysis and commentary technology in the chicago tribune, the washington post and on and on. the website that he has been a part of including passing reforms to the free information act and the june. weighing in on the foreign intelligence act.
5:00 am
mass surveillance programs. my thanks and welcome to all of you. arthur i like to begin by having you tell us how you developed this obsession with drones. [laughter] i'd like to think the institute for having me and i feel tremendously honored to be here on such a memorable panel. it really does me a lot of good to be back in the state. this is been a quite incredible journey of anything about the fact that not seven years ago, i was a pretty scrappy undergraduate in college in upstate new york. every morning i would read the times and the breakfast cafeteria and invariably there would be a story about drone strikes in undeclared war zones. if not that, there would be a story about how drones were increasingly being used in the
5:01 am
domestic civilian airspace. both of which raised unfamiliar and urgent questions. now for my part, i was doing my research with history students about immigrations in northern new jersey in the 1960s. i was sitting in a bar one day between my junior and senior year, and suddenly i had an idea. i have to study drones. i have to create something called the center for the study of the drones. i returned to the college and i told the administration and until the correct members we must do this and because they are likely completely insane, they allowed me to go forward with this. we created this little research experiment and the rest so to speak, is history. i guess our timing was fortunate because we established ourselves at the time when people began to ask these questions in a very broad public forum. those questions have only become
5:02 am
more complex and more challenging and more urgent as time has gone on. and so, i spent my time at cia as an analyst essentially during the very tail end of the cold war. into the mid- 90s essentially. >> i was used to working with both breather systems. also very highly classified satellite imagery programs some of which i can talk about and a lot i cannot even though it's been nearly 25 years since i've actually been active in doing any of that. but, what you say in the book about this whole issue of the soda straw in terms of trying to actually see something from above, then applies to pretty much any kind of conventional imaging platform including even relatively advanced satellites. the things that i that are that in the unclassified now are things like digital gloves
5:03 am
satellites. these things all operate and what we call the electrical optical spectrum. as a spectrum that our eyes use on a daily basis to see each other and to see the world around us. there are other spectrums of course that are of great interest of course as far as the law enforcement military which includes infrared and wave infrared. i wanted talked about some of that in more detail later. what a find kind of terrifying about this is we are now out of the you take a picture here and you take a picture there, you're talking about this whammy technology. tell us what that acronym w ami stands for. what does it actually mean a real terms. >> well i should say that i'm a drone researcher and i spent a lot of time thinking about these frightening technologies but in a way nothing kept me up at night the way this technology did. this levels tear is what it's
5:04 am
called. so as pat was mentioning, over the course of the cold war, this sort of pinnacle area defense the satellite that take images. once you move more into anti- arabism paradigm. you can focus in on individual people you want a video camera. but the aerial surveillance videos session them which are by a margin use, operate on what's called live sort of the soda straw printable. think of them as telescopes, they are very good and watching a very narrow area and a very high fidelity. but if something happens outside of the area that you're looking at, well you are out of luck. an example was given to me by one source says the air force and several agencies were tracking a senior and first leader who was in a convoy of edo they co- cults. he was differently in the
5:05 am
vehicle but they didn't know which vehicle precisely he was in a certain., they reach an intersection and they split up at that moment these alice had to make this very difficult decision. to be a left do we go right. it basically went on the flip of a coin. what if you could watch the whole area at the same time. that is the principle behind what i read about in this book whammy. this technology that cost me so many hours of sleep. where you basically get a giant camera, and you watch an entire city at once. the idea being that you can follow thousands of vehicles and even if you don't see the vehicle and the thing of interest in real time, you always have the footage to view later. this sort of genesis i should note of this technology is actually from the move enemy at the stake. it's from 1998, with will smith, is about this road sell within
5:06 am
this national security agency which will smith has some evidence that they want. they deploy a whole dazzling array of technologies, they put trackers in his pants and his shoes, they put a camera in his smoke detector. without a doubt, the most terrifying technology is the surveillance satellite which is able to see the entire eastern seaboard all at once and it hamza video capability and watches the will smith character as he stumbles around dc. seeing the satellite and operation is truly truly terrifying. as far as anybody knows, it didn't even exist at the time. but one night at a movie theater in 1998, an engineer working at the lab, went to see the movie, with his wife and where is everyone else in the audience was no doubt terrified with us
5:07 am
on the screen, he was absolutely thrilled. he thought it was amazing and he thought we should do this. so he rushed home and left a message with his supervisor saying something reasonable, i have a great idea, tommy. and so the scrappy teen that work of some ideas and have digital surveillance could be used in that capacity, ultimately they stuck some cameras together and it was all pretty scrappy, but they were able to watch a very large area. in the cia got involved and became very interested in what they were doing because they could use it to unravel networks of insurgents. in iraq where these networks were really wreaking habit in their service members with ambushes and ied attacks. you have a very wide area view, it doesn't matter if you don't see the id go off at the very moment. you can rewind that moment in time and see where the people planted that id came from.
5:08 am
none of that, you can see where they went but it gets better. once you see where they went, now you have a location associated with this insurgent group so then you can track all of the other cars that came to that location. in two other locations in theory you can find the people who like the really big decisions in this group. this is obviously thrilling to the cia because they were really trying to find a way to identify these groups that were essentially look like any other civilians. so they put this technology together and it went through as incredibly rapid series of it development cycles culminating with system that races cover in my book global stair, which is therefore system which is continues to be in use this very day operating in at least as far as we know, afghanistan, and syria and congressional reports act absolutely crucial capability. the thing about it is it is
5:09 am
classified but it is made a tremendous difference in that original role. >> at least as the claim that is being made on behalf of your sources right. >> yes. i think what we learn from the history of surveillance programs in the united states over the course of the last and most 100 years now, is that often times these claims of efficacy don't necessarily and out. an example of course would be the patriot act 215 telephone program which more commonly known as the call detail record program. even though that program was exposed, stopping exactly zero attacks in the united states in 2015, the congress went ahead and reauthorized the program anyway. i think that's one of the things that concerns me about not just this technology but a lot of technology that is out there right now whether we talk about facial recognition and other forms of biometrics and things
5:10 am
of that nature. these programs have a really nasty habit of getting funded and taking off and developing a lot on their own and never getting really the scrutiny that they need. to the best of your knowledge, has any inspector general either within the department of defense or any of the service inspector general's ever taken a look at any of these programs to see if the claims actually match reality? >> they certainly have, the technology is an uphill battle, there were a lot of steps and skeptics lot of people who said if you get one megapixel camera with a predator why would anybody need more than that. there was also some very scathing development testing evaluation data that came out about some of these programs. also, there is some evidence that the technology has as you said, escaped beyond its original constraints sort of set
5:11 am
of uses. one, senior officer who was involved on the analysis and, of the golden step program, said that it's been useful in a narcotics operation in afghanistan. that had nothing to do with what the cia initially intended for the technology but once it's there in battle, is those checks are necessarily applied. you use the tools that are at your disposal. so, that being said, i feel like the budget date is sort of speaks for itself. there are numerous ongoing programs, the army has new programs that develop some capabilities better than the marine corps', air force is continually investing more into technology and one gets the sense that probably has something to do with the fact
5:12 am
that it has shown at the very least tremendous potential in one forum or another. there is one data that i was able to get about these operations. there was one system, a set of four aircraft or blue devils, wide area cameras. according to one documentation, and a three year time span, it was credited, credited with the capture or killing of more than 1000200 people. in afghanistan. 1200 people. that is a very tiny pick in to what exists behind the curtain. >> you just reference kind of the use of the technology in a narcotics faction. let's make sure as we can be as fair as we can with respect to the technology. technology can ultimate be used
5:13 am
for good or evil as we have seen. while the same equipment that's used in pharmaceuticals can be made to make nerve gas. >> there is a flipside to the story too. i think it's important that we talk about that. i prefer to talk about it upfront then when we are pressed for time at the end. let's take a hypothetical here. if google had its own capability here, how much better would google maps be and how much better would your traffic management control system be if you are able to kind of employ this technology #. >> absolutely, i interviewed one official rather a senior executive at sierra nevada which is the contractor the prime contractor for golden stair, he was actually driving in dc while i was speaking to him. obviously the traffic was incredibly bad and i did a
5:14 am
little background on this in the technology can be used to identify points in real-time. it can be used to gather data to traffic models to how to best optimize the traffic and city. how to space your time in traffic lights for example. there is more to that. about a year before i started working on the book, i was writing my bike home from a bar. in brooklyn. i witnessed a shooting. it was a shooting of four peop people, they were shot at 19 -year-old in the gut. they disappeared into the night and obviously i didn't go chasing after them. i contacted the police the next day and i was in touch with the detectives trying to give the information that i had and i checked in public or so later and they were never able to solve that crime. virtually the teenager survived
5:15 am
but it joined this list of thousands of unsolved crimes in new york city. every year. had this camera been watching that night, it would been a very simple question of tracking, back in time to where they came from and also forward in time to where they ended up hiding out. even if that hadn't allowed the police to catch up with them every night, he would've given them an address. i want this people to be brought to justice. i saw this teenager lying on the ground. if we have the capacity to do so in a way it's kind of encampment to at least make use of it. but the story is never so simple, is it. i also heard about some very terrifying things it can be done with the technology in a domestic setting which thus make no mistake it's happening.
5:16 am
it is being used and there are groups that are trying to have it be used in domestic setting. they are testing in baltimore and a lot of other cities and just last week though as i referred to as the henry ford of this technology, announced that he now has insight on a place in chicago, to have the technology fly over the cities st. louis and congo incredibly large area to solve unsolvable crimes. and, the last thing i'll say about that is that it is completely legal. as far as the law is concerned, there is no difference between this man droning an entire city within a hundred and 92 megapixel military grade camera and me sticking my camera out of the window of an airplane to take a picture of the landscape. it's a public space and i have the right to do so.
5:17 am
>> sean, the attorney due by that question. >> i think that is fair to say that the law in this book that there is another's part of this that you are not god into yet. you call it abi at one point is a technical term. we are really talking about the artificial intelligence apparatus around it and the other similarities around it. the increased collection ability generates just way too much information for the normal intake process. i have a specific question on the other side of that it does start to go to the site but i think we will need to have you explain a little bit more about that before we get there. >> one of these cameras, single one, generates an unfathomable amount of data. i calculated that it would take like 2000 ipads to play the
5:18 am
imagery from a single camera frame at any given time if you're looking at sort of real-time of real size resolution. as one engineer put it to me, it takes a million people to watch a million people. and sure enough, when the air force began analyzing all of this footage, they found themselves completely overwhelmed. a vast majority of it was ending up on the cutting room floor. they can obviously find what happened after an explosion that was known about, but they were unable to find what rumsfeld refers to unknown unknowns. surely there are so many other things happening in the footage but they simply didn't have time to get to. so the solution to that is artificial intelligence. not only does it spare you the the bloodwork of tracking the
5:19 am
vehicle. you just say to the algorithm and theory tell me everywhere this vehicle is bent and everywhere it's going. you can also say every other vehicle that it is associated with, give me a list of every location and then track all of the vehicles that have been to those locations as well. but there is more, then want to start with someone who is a known terrorist, maybe you want to get as the pentagon would put it someone with a list of a bang, as it turns out, these groups often exhibit some pretty predictable behaviors and laid up towards an attack. they will do some pretty simple surveillance so they will take new turns or they will drive aimlessly so that they know no one is following them. but what if you told artificial intelligence to tell me every time the card exhibits one of these behaviors in a city. now, the system even if it
5:20 am
doesn't catch everything a one, it was catch some number of unknown unknowns. and that, is a true holy grail of surveillance. to find everything that happens that you had no other way of knowing about. there has been an intense effort in the last few years to automate this technology and a lot of people no doubt have heard about project maiman is a controversial effort that google has been involved in in other silicon valley firms. it's an experiment sort of gave some automated capabilities to some footage. now it has turned to wide-area emotion and injury. so that has received far los attention and it is something that we shall definitely be thinking about. i'm not sure we want to live in a city where every time someone does a u-turn, it seems a little suspicious. they all of a sudden have a cross. i'm.
5:21 am
>> i want to get that out there because from a legal side, the question whether or not this is legal in part as well as how to use it and why. we have and you talk about this and to some extent in this book. that generally speaking in the population there is an upper bound to what this would look like. so, there is a private application but the idea of introducing this at trial and using it as a true technique it domestically. i'm not so sure that i would agree with whether it is legal or not. it is unaccounted for perhaps. >> i have challenged unchallenged. >> so here's the thing the jones decision. that john is referring to in 2012, it involves police use of a gps tracking device on the subject's vehicle, not for a day
5:22 am
or two before something on the order of weeks. >> i believe it's 20 days. >> see you are talking about essentially, placing a specific device on the vehicle. and having that person tracked literally in this case for roughly a month. the court said no, that is basically a violation of the fourth amendment. for me the question is is inapplicable here even without the application of the actual device on the subject vehicle because you are literally now utilizing a very different forum and persistence surveillance are just not sticking the actual receiver if you will, whatever you want to call it, the tag on the car. so it does make me wonder whether or not jones would potentially be operative here. is there is no actual case that you are aware of in any federal jurisdiction right
5:23 am
now or this is actually and isn't this part of the reason why that may be the case that as well as the case in baltimore persistence surveillance systems, they did everything they could in baltimore pd to keep that use of that system absolutely secret. they did not want the public to know about it. some had no knowledge of it led to community outrage. isn't this penchant for secrecy and using a designation law enforcement sensitive information, isn't that essentially why we haven't seen a challenge to this kind of thing. >> without a doubt. if you think about how recently congress has began to pay attention to cell site to track ourselves. largely attributed to the fact that the fbi allowed a local law enforcement agency to use the technology.
5:24 am
nondisclosure agreement. you're absolutely right that in baltimore it was a secret operation and the reason it was secret was because it wasn't funded by the city. there was a texas billionaire philanthropist by the name of john arnold who actually gave the city enough money to run this program to see if the technology did have a potential. as a result of that loophole, the police department did not tell the mayor, it did not tell the state legislature, not the city council, not the public defender, the list goes on. i was lucky enough however to find out about this operation. while it was happening. i was sworn to secrecy and i spent two days in baltimore deserving the city along with these analysts and their departments the police department it was incredible what this technology could do. i sat in i should say on a briefing for three detectives who are working on a murder investigation of the shooting that was very similar to the shooting that eyewitness.
5:25 am
the analyst showed how they had been able to track this man's assailant for hours. it was following the shooting and how it led up to it. . . . i knew the airplane was watching. i looked up into the sky and i could see it flying high. it felt pretty uncomfortable to know i was being watched but far more uncomfortable i will tell you is seen everybody else going
5:26 am
about their business in the city. knowing that they are being watched by a technology that they cannot even fathom it don't even know exists and they have no idea. to me, that felt wrong and fundamentally wrong that technology is watching everybody and nobody knew about it. it is wrong on a moral level in that sense and a visceral level and it's also wrong in terms of the fact that the city councils do not have an opportunity to weigh in on it in the states attorney did not get an opportunity to way and whether this evidence would be admissible in court and it ended up having having a whiplash the fact because when it was finally revealed there was so much outrage but [inaudible] >> as you are reading the book what struck you about the
5:27 am
journalistic techniques and his ability to get these people to talk to him and i was floored that people who had worked at some of the most super labs in our country and worked on supersecret stuff at my former employer, cia and other agencies as well were eager to talk -- >> absolutely. it's funny when you report on this intersection between technology and intelligence communities sometimes you happen upon these topics that people are excited to talk about. they worked on these things and developed these new godlike tools and they want to brag about them to a certain extent. within the intelligence community particularly within the research labs there is this space that exists pre- classification if you will the people are scientists and may have a different frame of mind when discussing these things. i would like to hear more about your process of reporting and astounded by the amount of information in there but i also,
5:28 am
on top of that, would love to bring up the ai issue as well as a journalist who is covered the intersection between technology and the intelligence committee something i'm often struck by is the reliance on this godlike tool and the tendency in which in its early stages is very prone to [inaudible] i wrote about the cis using corporate technology to connect resources on the ground in china, iran and essentially it's a webpage media source that into yoga. it's a webpage they go to and say it looks like they're bragging about yoga but in reality they are communicated with the agency and an overreliance on the tool that is not that secured led to the deaths of sources around the globe so i feel like looking at the errors of that particularly ai [inaudible] super dangerous.
5:29 am
wearable computers make mistakes? when you look at the algorithms that people talk about how dumb they are at this current stage and were rushing headlong into something were not really prepared for. >> i will start with the final question and coming to jenna to state my reporting was up for a tremendously high praise and i would totally get that tattooed on my arm. [laughter] it was funny when i started researching the book i would tell people this community journalist and navigate that i was writing a book about [inaudible] and barely people would say this guy will talk to you and i started to get nervous and said but as it turned out it was a story that you referred to where people were willing to speak to me. there were a couple of reasons for that, i think. one was that they are proud of what they've done and the core
5:30 am
development series of this technology with powers of the camera outpace the growth of my computing power which is described by [inaudible] that's an astounding achievement if you think about it because were talking about the periods of the early 2000, five years ago, computing really improved. we wanted to talk about that and brag a little but much more important i think is that they had a sense that it's an important story. this was a story of broad public interest and that they almost have some obligation to get it out there. in fact, one of my sources finally reached him and he said arthur, in a way i've been waiting for this call for 15 years. i said why? he said because we have to answer for what we've done. they have had a sense from the
5:31 am
very outset of this program that they were creating something formidable and there was one incredible moment right in the happiest moment of the development cycle where they felt like every single day was the day that they were not saving u.s. servicemen in the field [inaudible] everything broke down. they were in florida, during surveillance in palm beach without anyone knowing about it. then we all went down to the beach and apparently they sat on the beach and were a little overwhelmed by the enormity of what they've done that perhaps they set off a process that they would no longer have control over and another thing that really amazed me about the discussions i have had -- in every single one i intended eventually to bring up the question of privacy even if
5:32 am
these guys are talking about strained set of military uses i wanted them to talk about privacy and without exception every single one brought up privacy before i had a chance. they were thinking about it, too. this gives you a sense they're all in this discussion and it's a crucial role they play in this public dialogue but also the technology itself does raise these difficult questions. to your point about artificial intelligence, you are right, it's [inaudible]. one of my favorite anecdotes was group at colorado state developing this recognition algorithm for having video surveillance and the system was very impressive and been able to detect a woman and that she was turning around but as research but it admits that she was
5:33 am
carrying a bowl of fruit. you can't trust the system that misses a woman carrying a bowl of fruit and i thought it was kind of strange prop to use but there were a couple of things that play there. they don't care if they miss even a large number of the suspicious activity. as long as they capture more than they would be able to catch by human analysis alone and that's a compelling idea. the also the fact that they know the stuff will get way more capable over time. they were all just starting to use this technology called the deep learning. they found that it had tremendous potential in this application but your question raises what i think is the key issue with regard to all the surveillance which is the question of trust. not every analysis by the computer is going to be based on
5:34 am
fully robust data and so the computer needs to give you a sense of how much it trust its own analysis. it needs to save you this is 99.9% likely to be true or it needs to tell you i'm not really sure. if you think about it in a law-enforcement context that is incredibly problematic because it says to you i have a 71% confidence rate that there is about to be an armed robbery. do you send the police and with their guns drawn to mac do you meander over with a single patrol car? but if the computer gets it right and you did not pay attention. next time it gives you a 71% confidence and will send in a whole swat team. what if it has a high confidence rate is 95% and it was some teenager playing ball on the street. the next time it gives you a 95% confidence analysis you will not trust it at all.
5:35 am
it will not be long before a lot of police departments start to have very dysfunctional relationships with their ai and that is not only it's problematic because it needs the effectiveness of the technology will be problematic but also problematic because people could really get hurt. >> there's another layer below that as well and it's a fascinating part but it's not just that 71% -- which is certainly true but i will be two passages that stood out to me. as a matter of policy must be beers for one of the software involved here track vehicle events of interest and included in those are [inaudible] keep distance close, keep this in far, flip-flop driving once cars pass each other repeatedly, approach, retreat, parallel driving, dropping off,
5:36 am
illustrating and meeting. presumably those factor of things like that into that 71 number or some analog to it and the more alarming goes back to the legality questions that i think it's incredibly important thing for research and building not record here that additional markers identified by secret pentagon reports for anyone -- if there's one out in the next 24 hours that those include name, gender, age, weight, religion, skills, biometrics, values, e-mail addresses -- >> red flags. [laughter] >> how this question is framed is mildly important. curious about generally speaking where you fall in the broader spectrum of the evangelist for this type of work type and maybe
5:37 am
the more privacy [inaudible] just the granularity there inspires somebody just not just new questions but philosophical questions. these are qualities that even in a war zone i think would inspire some pause in triggering life or death situations. >> is that what are they doing about that? you get all these great interviews in this book about from the evangelists who have [inaudible] let's do it all the way to i had to stop working on this. it's not clear that talking about these but they recommend issues you bring up probably so how do we address transparency but these are fundamental issues and i also -- honestly i make the concession that it works in absence. >> it touches on an important
5:38 am
point which is that the reason this technology is significant is not fully because it's powerful in a vacuum or on its own but because it is emerging at a time when we have the capability to find out a person's name and gender and religious beliefs and their associations using things like social media analysis and wildest medication and substance so when you bring all this information together and then when you apply big data, analytics to it to find that a person is not only doing a suspicious set of u-turns in the middle of the night but they've also posted information on facebook that shows they have a particular meaning politically meaning that may be of interest. that is immensely powerful. that is tremendously not only because it has the granularity element but also because it touches on the automation
5:39 am
element in the idea of [inaudible] and also because it raises these questions of whether it leaves anything that can remain private in this day and age. there was this really incredible passage that i found in one document talking about how a system that would bring all the separation together could be used to find osama bin laden. it does so by doing exactly the things and i did find suspicious driver behavior and so on and so forth. the answer that these folks would give you to your broader question is that they have a single mission in mind and have a single cost and will stop at nothing to do it. they are absolutely dedicated to saving people's lives. in a way i came to realize after initially being frustrated by that answer but that's the role they play in this ecosystem. you can't expect the lion to
5:40 am
decide it suddenly not going to eat eggs out because there aren't any gazelles left anymore. these folks are in this technology space and they will do everything they can and yet, they have some sense that something needs to be done and they will suggest what needs to be done about it but in a way there are others in the ecosystem that they need to respond to an spoke to an earlier question about it's within this intelligence committee and my answer is no not particularly but we talk about drones must agree there's privacy for it but i don't think it's necessarily on far enough. in a way that's why i wrote the book and that's why you guys are willing to speak because they've done what they did and maybe
5:41 am
that happens maybe they're not and now it's our problem to deal with in a way and i would not put those words exactly in there now but that's the situation. the cat is very much out of the bag. not only is it out of the bag but it's turning into this frightening tire with many eyes and ears listening to us, automated duties frightening things and it's miniaturizing now a camera that used to way a thousand pounds weighs 30 pounds and you put it on the drone with one engineer i spoke to cap later that 72 drones that in the 30-pound canvas and watch the entire island of manhattan. not only watch the island of melinda manhattan because the low find drones as opposed to one single high flying aircraft around buildings. you can build a 3d model so that now the police would be able to have a 3d explorable environment
5:42 am
of the city. imagine a really realistic videogame of manhattan where people are moving around but those are real people. you can fly around in this virtual not virtual 3d space. that's where we are headed. and these guys are not going to pause. >> talking about this level of granularity in the data one sort of popular vector or conversation recently is what happens if an understudy deals that information connect you talk about how they're protecting it? constantly we charge into the spaces and collected all but we don't actually protect it and then well, weight, china has it all now and they're adding it to the opm data and equifax data and are people thinking about that? >> very much so. the answer you will get is that trust them. we secure this data and i'm not
5:43 am
a cyber security expert so i don't ask them to show me the code for using to secure these systems but that is another huge question that the scenario i imagine is that you have a city with a very well regulated surveillance system and in this city on one day system is updating the happens to be a black lives matter protest and a hacker associated with the white premises group gets access to that data. now they can track every single protester who attended that protest back to their home. >> the entire city of baltimore so it's not theoretical. >> to your other points, we assume that if this -- this continues to be employed domestically -- international terrorism and clandestine use and those are the for instance
5:44 am
for surveillance that is the stuff the government lowers standards for accessing [inaudible] the specific way they use the call veto records is [inaudible] and it's not just the one person you're looking at but it's everybody and all that [inaudible] by the time you get 2 degrees out which is the law under section 15 you are talking about a program in this case would be under constant monitoring and you're talking about surveilling people who not only are protected of doing anything but the vast majority of whom have not been in contact with those who are trying to follow. that's so much different than what we have. >> one of the tactics of the baltimore police permit experiment in with is not just to track people who have been
5:45 am
involved in the shootings but people who have yet to be witnesses so they can track them back to their home addresses and have two police officers knocked on their door to extract testimony. we all have a right to withhold testimony if we don't want to and if it's just the police officer casually dropping by our house intimidating that may be this is a way to cnet networks. that is, by definition what it is for and we are so interconnected as was mentioned in these different ways that the only possible sort of affects i can imagine from this is this chilling effect. now that i know a surveillance system will potentially put across their army if i drive through on the neighborhood that they has a high conversation of people who appear to be involved in the drug trade because it determines that i parallel
5:46 am
parked next to the car or have an association with a person i will not go to that neighborhood anymore. it's interesting i found this one report that talked about bowls of wide-area surveillance and its -- with wide area surveillance the joint operating manual we want to give -- this is pretty much a )-right-parenthesis price we want to give the adversary the sense we can perceive that even are very intense and not just what they're doing but with a plan to do. we want them to look constantly over their shoulder even if or not directly watching them. it's well and good if you delete with the violent terrorists but the technology will have the exact same effect domestically. it doesn't help that is the gorgon stair. it's very much intentional and all these systems have very formidable names, civilian
5:47 am
gorgon stair which is not been as far as we know deployed domestically and active operations although we know that fbi and department of homeland security have taken a great interest in it but it's called vigilance there. it's tony it down very much but it's also constant or blue devil as i mentioned, angelfire, there's a lot of greek mythology there. >> its history teaches them anything it changes it to something boring but they get scrutinized. >> but it further doesn't help other things we touched on is that the programs are predicated on race. in religion. you know, i'm curious when you interview [inaudible] they talk about transparency and as someone who supports this supports transparency along these lines for what they doing
5:48 am
to get to these fundamental questions? is there -- do you see a way for them to strip out some of these were obviously problematic components, i suppose, such as it would be something that would be palpable. you clearly see doctors for use domestically and i don't see how that could possibly happen in the absence of tackling these -- >> absolutely. there is no doubt that there needs to be stricter controls. the company has a privacy policy so they can go above and beyond what's required of them and there's pretty sensible rules in there and things like there will be a meticulous log of every action by the analysts who access the imagery so that no analysts stopped following their spouse for example and they create minimum operating attitudes to ensure the technology does not have
5:49 am
resolution to identify people's faces. in a way it's not these folks response ability to hold themselves to account nor is a particularly wise because they are subjective at the end of the day. ross said to me i will not be political surveillance but in this was literally his next sentence i will do a world bank type of thing where you have 20 agitators anarchist trying to make trouble. if you look at the historical record there are some very fine people have been referred to as agitators. think about how the civil rights movement was spoken about at the time. there was another really chilling story and another engineer who developed this technology and is trying to commercialize it said that he used it one day to fly over the michael brown protest in st. louis and spoke following
5:50 am
the shooting of this black teenager. he said it was simply a test that he wanted to test out some algorithms against the large crowd so that people on the ground were test subjects. >> yes, test subjects. >> and he also said from the conversation he vehemently proposed the black lives protest and even refer to some of the protesters as thugs. i asked him fair enough, you weren't breaking laws but if you saw anything that raised concern would you have told the police? he said, of course. that's my responsibility. but what is his threshold and what counts is concerning activity? he says i think this pixel on the screen blew a brick through
5:51 am
a patrol cause window to get his address so what? if that capability had existed during the summer montgomery march who knows where we would be today? those are the same rationales that we use against those groups. i have no reason to doubt that these guys and i should be sensitive to the element, by and large most of the people i spoke to were men and the only person i spoke to who said i would not want this fight over my own backyard was one of the women. >> with a mostly white as well? >> yes. mostly white. i think is relevant that the engineer is white and his wife, however, who vehemently opposed the black-white matter protest movement is black. i was thoroughly confused by
5:52 am
that. there is that subject to the element that was very palpable in all of this, even though they had the best intentions. they wanted to have peaceful society but we all have different ideas as how to get there and perhaps the people with the access to that analogy are not should be calling the shots. >> within the law-enforcement committee within the national security community especially it's an extremely mission oriented culture that the issue is how is that mission to find, right? the mission is always get the bad guy, get the bad guy. that is the message the public, politicians, they sent him pretty much a regular basis.
5:53 am
that's why i think a lot of skepticism essentially has to be applied to the claims and the advocates at page 180 with the passage got to [inaudible] he was on the majority side, republican side on the committee for intelligence and this is what he said in response to some of their issues you are raising. the fact of the matter is he repeated a line [inaudible] u.s. intelligence community and ability to spy on you but does not because and here's the secret, listen closely, he paused for effect u.s. intelligence community does not care about you. mike, if you are watching, buddy, you might want to
5:54 am
rephrase that. might want to go back and rethink about that. when mike was on the hill is when senator chuck grassley revealed in 2012 that the nsa respect the general had found nsa employees were misusing nsa systems to listen and on conversations of their current spouses or former spouses or former lovers. this was one of two people but at least a dozen in those were the ones they caught. i think for me because i've been in this town for so long and worked both in the intelligence community but also on the hill and like caps on i've been around for a lot of these surveillance battles over the last 15 years what always concerns me is the back and process essentially trying to keep tabs on the stuff. you can pass a law and this is what the foreign intelligence surveillance act was designed to
5:55 am
do one passed in 1978. ... what gives you reason to believe that we have a better than even shot of being able to get it right with this kind of technology because in many respects, kind of like you, i am fascinated by the technology. i love this book. it's a superbly written and beautifully edited.
5:56 am
but i know the mentality that drives an awful lot of folks in the intelligence community and i worry especially about the mentality of the folks on capitol hill who seem entirely too eager and willing to let things slide. we have this month just last week another example of congress not doing its job. to the extent that you can.
5:57 am
at least at the level. there is warrantless wiretappi wiretapping. but that question is that we really have no choice. you know, you would think having spent the better part of four years working on this book with all of these frightening technologies, i find it hard to sleep at night and i would be barely pessimistic about the prospects. but i don't see that as being an
5:58 am
option. i don't think people who are pessimistic go to the council meeting or write their congressmen. i don't think the people who are pessimistic challenge the technology in the courts. so, one of the saving graces about this technology in particular is that it is palpably scary. it isn't any sort of abstract things. you talk to people about the data analytics and it's a little hard to wrap one's head around it but when you tell someone that there is an ie in the sky watching them, that make sense on a more intuitive level and i think the reason history speaks to itself in that regard every time so far a city has revealed that it either has this
5:59 am
technology or is intending to use it, it is being derailed by the public pushback. so perhaps we have reached a technology that is 124 where people say wait a second, we care about our privacy, tha bute are not going to be watched from the sky persistently. that is just not going to happen. i think we can use that to our advantage. to date there hasn't been a single congressional theory about wide area surveillance technology. no public hearings that we know of. we don't know whether or not the house and senate intelligence committee has held closed hearings and just to jump in real quick as kind of a historical note of a document from house or intelligence committee has ever been transferred to the national archives. that's another area where we need some more transparency. i just have to get that plug in because it is important to our ability to see whether or not the committee in obvious ways they are not doing their jobs.
6:00 am
but when the committees themselves are in charge of overseeing the programs they are not being transparent in the way essentially the constitution requires them to be. we should have some concern that the state and local level, you see hope. >> it is easy to talk about it and explain what this does and why it could be beneficial. my hope is that that begins with a level of public awareness. that is one of the main motivating factors behind this book, and i think that when people know about it they have a desire to take action but there's an important caveat to that which is more enforcement by nature will always operate
6:01 am
according to what the rules say and not according to what they do not say. and i will give you an example of that. let's say we create a role that is very much robust and creates all the limitations we could possibly imagine on aerial versions of the technology. and the police department on top of a skyscraper. this is a ground-based system so the regulations of the technology take into context all of its iterations. it has to be future proof which the only solution two is thinking about this as a process rather than a single roll of developing a golden regulation that will be completely watertight forever and that there is an emphasis on making sure the rules that comply to as
6:02 am
they are written but also as to what they do not say as well. and i should also say because we have seen cities take these actions, they just passed a very robust municipal ordinance that requires the city to disclose any new surveillance technologies for the baltimore observation wouldn't have happened and to submit the technology to a thorough review and regular orders to make sure no one gets lazy in enforcing i think that is a pretty solid step forward. but perhaps for that to happen at a larger scale i think a lot of people are going to need to participate.
6:03 am
a lot of my sources are a little bit alienated by seeing us here on stage talking about the tension and how evil this technology is. one of the issues has been coming and i'm sure that it implies an orders of magnitude here is the cost of storing this stuff. how long do you keep it, what is the public access process for it etc.. talk if you can about the magnitude of the storage
6:04 am
challenge. >> it's absolutely huge in a way that technology also creates interesting loopholes it isn't relevant to the ongoing investigation. it has to be deleted within 30 days. the data storage is becoming a lot cheaper. every day you call every third
6:05 am
frame of its annual meeting of one third third of the storage and then you automated and compress it and there are technological fixes that need to be considered. >> the state and local police departments moving away from helicopters because of the cost and moving towards drones do you see a movement like that accelerating and potentially being a factor that accelerates the adoption of the technology we are talking about? >> now the police department in the town with 300 people put an eye in the sky if they could previously do. there are the most recent colleagues more than 900 public safety departments in the country that operate drums. these are very small drones. but there are a few things that
6:06 am
are converging. they are opening up for and more to the large drones that would be able to carry something like this and eventually you get to the point that you have these that can watch a whole area persistently and actively and intelligently and that will be a few more years down the line to solve all these problems and it's going to see everything and catch every bad thing before it happens. that's just immediately thinking, i can think of all these different problems it might have, for example, north korea. it's very cloudy and the terrain is very mountainous. i heard a lot of imagery and how
6:07 am
that is a difficult target because of the various issues. you could have people shooting down drums. you think that people have thought about these issues that are starting to counter that and on the other side, do you feel like it is being built in a way that gets them a lot of money that mighbut might leave out the potential issues that they've run into? >> mike gave a good answer to that. he said our job never ends. we can develop the best system ever to seem like it will watch everything, but our job is not done. we will continue to do what we do and it will be better and bigger if we want to see this if
6:08 am
we want to see the class we will see through the class. the system could pick up very faint traces and extrapolate those full views and now maybe you don't use a camera to synthetic aperture radar. these are all minor roadblocks in the way no one is throwing up their arms. and if another system comes along that is going to achieve the same goals more effectively than money will get pushed aside. i should say no one is referring to it as a cure-all come as the only game in town.
6:09 am
a lot of the systems in use today have other surveillance aboard an sublet one could detect something happening in another part of town it could take a closer look were to track every cell phone in the area or accuse radar to see through the cloud cover. it's when all that stuff comes together that i think you get to that kind of spiral of viewing. >> so we kind of started this with enemy of the state. before we go to the q-and-a to kind of end with this and the whole idea of a satellite-based capability. you spent some time talking about that. but walk us through your understanding of where things stand in that development. >> as we know from the recent
6:10 am
events, they are very affordable to the antiaircraft systems. we are approaching a time when the war that had unfettered access where we don't have that privilege and so in theory a simple solution to that is to put these technologies into space and what i found if there is a convergence of a number of technological trendlines that take us in that direction whereas a satellite used to cost half a billion dollars in cost enormous amounts of money and resources to launch into space and now for less than $100,000 once the imaging technology is small enough to put now it can collect wide area video. okay, while satellites travel
6:11 am
with 17,000 miles an hour and get about 90 seconds over a target for the crossover. why not just for the whole bunch of them so the second one is able to talk over the other side to pick up the slack. you start to get to a point you can see things pretty consistently. in the early time in what people refer to as a space revolution. the idea of the wide area surveillance with large.
6:12 am
unblinkingly but movingly. i took this with a little bit of a pinch of salt with somebody saying we all have flying cars in a couple of years. but then a couplwithin a coupler that conversation, a company called earth now, not only that, but they have financing from bill gates and a number of other investments. you don't take bets on totally ludicrous ideas. they didn't say who they were going to allow to use it or what they were going to use it for. we still have an information void when it comes to us now. they've got their eye perhaps on it. >> wait for the microphones to come to you. and then i will ask you to
6:13 am
identify yourself in affiliation and please do phrase it in the form of a question. >> right here in the striped polo. >> i want to ask you how we work from the permissible to the impermissible if we start with the police man on the street corner then we have four policemen at each intersection and then four policemen on each corner of the so-called bad neighborhood and then on every intersection of the city when do
6:14 am
we go from what isn't permitted to what is? >> it's a good question. it is part of the fundamental question all modern life is how we balance safety and privacy. some will tell you it is a very simple answer to the question. anything that makes us safer we should embrace because it will only be a problem for those of us who don't make them feel safer. i think we are now coming to the point we realized that actually, no, that is not a one-sided equation. look at all the cities around the country that have removed traffic light cameras.
6:15 am
they are a pretty simple technology. people said yes we understand that. the e. patient about the relationship with social media companies has changed. he said ithis is a great servic. i think that there could be a change happening.
6:16 am
>> often times when the technology develops, there are good applications. 56 years ago [inaudible] what are some of the positive applications that you see coming out of technology? >> great question, important question. in the aftermath of hurricane katrina, there were hundreds of people stranded on rooftops. they had to spend a great deal of time.
6:17 am
you would be able to find them much more quickly. you could watch a large section of the fire at once and you can anticipate before there is a lot of interest in using the technology to monitor the borders and pipelines for example. that can go either way. there is one company that has used the technology to find them in alaska to make sure oil and gas companies operating in the area and maintain a minimum distance to.
6:18 am
since the technology is so expensive is that having a single entity operating, you could have a company provide something like a google service where they put the information out there and anybody that wants it can buy in and that is predicated on the notion that there are potential applications we haven't even imagined. maybe you have a real estate company that wants to track them over the entrance companies who is to blame in a traffic accident. so, that is one of the models that has the proposed split of winning option. again though, we need to ask
6:19 am
ourselves whether we want to have google earth at any given time. i would use it to make sure the beaches were not crowded before i decide to go there but i'm sure there are things people coulcould conceive of thatcouldn conceived yet. >> the coast guard must be super interested. >> they've been testing a system that is incredible actually. it is a small camera that has a tremendously powerful algorithm is able to detect any object of the surface of the book that is not the water itself.
6:20 am
they took one of these drones about 50 hours to scan in detail in area. >> we have time for one more. let's go over here. >> i am curious if you think part of this type of activity and why wait until dusk.
6:21 am
that's an interesting question. i'm not familiar with that happening. at the moment the only groups that are using very large drone are the military and this was under very constrained circumstances so if they are not operating in a constructive airspace when it is because they have a single individual mission for something like search and rescue wildfire support. so again, i am not totally familiar with it. if you are talking about the capabilities and a navy is really only shows its value at night. which notably is not legal. >> that is a very good point.
6:22 am
>> it is the infrared camera technology to survey people not just from the sky but from ground level without a warrant because it is much more colorful, and because it has this standard of being in on publicly accessible as the supreme court would put it. that being said, the military grade camera doesn't to me sounds like a publicly accessible technology either so perhaps it is time as you mentioned the law began to catch up a little bit with the reality. that being said, we have seen
6:23 am
how originally it was only the governmengovernment itself thate resources to either drums because of their cost, and now any of us could basically go to a best buy or fill in the blank. >> which raises an interesting possibility. that would be the claims on the ground. the locations were later searched under the war in and found that the basis was actually dubious at best but it didn't exhibit patterns that resemble those of a site associated with the drug trade.
6:24 am
for all of you in the auditorium, thanks for coming out. a half.
6:25 am
6:26 am
6:27 am
>> i call this public forum to order. if


info Stream Only

Uploaded by TV Archive on