天空之眼：高尔贡凝视的秘密崛起以及它将如何监视我们所有人》，亚瑟-荷兰-米歇尔合著

JOANNE MYERS: Good evening. I'm Joanne Myers, director of Public Affairs programs, and on behalf of the Carnegie Council I'd like to thank you all for joining us.

Our speaker is Arthur Holland Michel. He is the co-director of the Center for the Study of the Drone at Bard College. We are very excited for his discussion on his recent book entitled Eyes in the Sky: The Secret Rise of Gorgon Stare and How It Will Watch Us All. In it, he discusses the future of high-tech aerial surveillance from its secret military origins to its growing use on American citizens.

Are we entering a new age of surveillance? According to our speaker, we are already there. It's no secret that surveillance has changed a lot in the years since 9/11, but with innovative wide-area motion imagery we have entered an era where, when seen from above, you can be tracked, your next move anticipated, and your last steps retraced. You cannot hide.

In Eyes in the Sky: The Secret Rise of Gorgon Stare and How It Will Watch Us All, Mr. Holland Michel tells us how the Pentagon developed a science fiction surveillance device that will someday be used in every major city on the planet, and it is called Gorgon Stare. This video-capture technology, which was developed by the United States military and inspired by the movie Enemy of the State, is a spherical array of nine cameras attached to an aerial drone. It is revolutionary intelligence, surveillance, and reconnaissance technology all in one. This drone and satellite technology reaches across foreign relations, politics, science, law enforcement, and privacy. It is Big Brother reinvented for the 21st century.

While Gorgon Stare is clearly very useful for the military, it has also proven to be of great value in isolating terrorists and bringing them to justice as well as protecting us from known and unknown dangers. Still, this protection compromises our privacy and has the potential to permanently alter many aspects of our society forever. The question, then, is how can we balance technology's limitless potential with its resulting impact on our daily lives? Is there a solution for preserving our civil liberties?

To explore these issues and more, please join me in giving a warm welcome to our guest today, Arthur Holland Michel. Thank you so much for joining us.

ARTHUR HOLLAND MICHEL: Thank you so much. I appreciate it.

Good evening, everybody. First, I would obviously like to thank the Carnegie Council for having me here tonight.

I actually have some history with the Carnegie Council. When I was fresh out of college and just starting the research institute where I work, the Center for the Study of the Drone, Joel and the team here were kind enough to allow us to co-host some events. I have always been very inspired by the work of the Carnegie Council, and to say that it is an honor to be here would be a vast understatement.

I'm here to talk about our overwatched future, but in order to do so I need to give you a little bit of history. Has anybody here seen the movie Enemy of the State? Okay, at least a couple. In 1998 a movie called Enemy of the State came out. In it a lawyer from Washington, DC, played by Will Smith, comes into possession of some evidence linking a senior official at the National Security Agency (NSA) to the murder of a congressman. This official obviously wants to get this evidence back from the lawyer.

What happens next is a frantic pursuit over the course of two hours of the movie, several days in movie time, and during that the NSA deploys a wild array of surveillance technologies to track Will Smith down. They hide miniature cameras in his smoke detector, they put trackers in his clothes and in his watch, they tap his phones, and they access his bank records. But without a doubt, the most formidable technology that the NSA uses in its pursuit of Will Smith is its surveillance satellite. This satellite, which created a key aesthetic for the movie, which was inarguably its most compelling feature, if you will, was able in the movie to watch the entire Eastern Seaboard at once it seemed. It could track Will Smith and his associates wherever they went.

The movie doesn't portray anything new in the sense of power trying to hold onto power and crush the weak, but what's new is that the surveillance technology accessible now—as the movie would have it—makes power all the more vicious and further rigs the contest between the weak and the strong.

On an evening in 1998 this movie was playing in a movie theater in Northern California. A man who works at a government laboratory, the Lawrence Livermore National Laboratory, attended with his wife on a date night. Whereas everybody else in the audience was no doubt mortified by what they were seeing on the screen, he saw this surveillance satellite and was thrilled. He thought, We must do that. So, as the credits rolled, he rushed home and left a message with his supervisor. It was very short. He just said, "I have a great idea. Call me."

This sparked a government effort to develop something that resembled the satellite in Enemy of the State. What this team did initially is they got a small number of digital video cameras and they strapped them together. They didn't have access to a satellite, so they put it on a helicopter, and they pointed it down at the Earth. With this imaging power they were able to capture a very, very large area, and in their early tests they showed that they were able to do exactly what the NSA had done in Enemy of the State. Their idea was not to illegally track lawyers who had come into possession of evidence linking officials to murders of elected officials, it was actually to monitor suspected nuclear sites in places like Syria. But they didn't have so much success in selling that concept to the government agencies.

Fast-forward a little bit to 2003, and it begins to be clear that the major threat that the United States now faces is the improvised explosive device (IED), and as one general put it in a classified memo, "If there is one thing that is going to completely derail U.S. efforts in Iraq, it will be these IED attacks." What he called for was a Manhattan Project-like effort to stem this violence.

Shortly after that, the Central Intelligence Agency (CIA) got put onto the case, and they enlisted the help of an organization that some of you may have heard of. It's the JASON defense advisory group. This is a very secretive group of civilian scientists who convene once a year to advise the government on a very complex national security issue. The government and the CIA said, "JASONs, we want you to look at IEDs."

In their first meeting they looked at technologies, how to put armor plating on Humvees, for example, how to neutralize an improvised explosive device from a distance, but they weren't convinced. In their second meeting they decided they were going to look at technologies for what they call "attacking the network" because as we know IEDs were planted by insurgent networks. These are groups of people who look no different to their surrounding civilian neighbors, who are spread out and diffused within a city. You want to find them because you don't just want to get the IED once it's in the ground, you want to find the person who put it there and maybe the person that person works for so as to stop the next attack from happening.

The official from the CIA who was organizing these meetings actually went on a visit to Lawrence Livermore and had a briefing about this Enemy of the State-style surveillance technology that they were working on. He was so impressed that he invited this engineer, John Marion, to attend his next meeting of the JASONs on the spot.

In the meeting there were a few technologies discussed, and then one of the civilian scientists, one of the JASONs, said, "I just can't bear this. If only there was a way to watch the whole city at once." Then the engineer stood up and said, "Well, I have exactly that," and he described how if you watch an entire city, sure enough an explosion would happen at some point in the frame. Once you have the explosion, you can zoom into the footage after the fact and see who planted that bomb and track them backward in time to wherever they came from. You can also track them forward in time, to where they went.

Once you know where they came from and where they went, you have locations that are associated with this terrorist network. Now you can track the other vehicles that went to and came to this location, and then you can track the locations that those vehicles subsequently went to. Soon enough you would have a full picture of your adversary's network. Obviously, to the CIA this was catnip, and so on the spot they decided that they were going to put their full weight behind this technology. The man who organized that meeting is still referred to in these circles as "The Godfather," because he put the people together to make this technology become a reality that we now contend with in modern life.

A couple of years later it was deployed to Iraq. Not long after that it was deployed to Afghanistan, and it was used in exactly the way that it had been described in that meeting. Vast tracts of the city were monitored, insurgents were tracked from the sites of incidents back to their safe houses, and the information of those locations would be sent to troops on the ground so that they could then in theory and in practice as well round up these individuals.

All the information about these operations is actually classified. Nobody was willing to tell me how many insurgents were captured, how many people were rounded up, and crucially how many U.S. service members' lives were saved as a result of all of the IEDs that were taken off the map. But we do have some hints as to the effectiveness of these systems.

Soon after the first system was deployed the Pentagon began investing in a whole range of aircraft built around this same concept. One single aircraft that they developed, called Blue Devil, was deployed to Afghanistan. It was a fleet of just four aircraft, and in three years it was credited with the killing or capture of 1,200 suspected insurgents—four aircraft. That is a staggering number, and that is just one of the systems that was in use.

But they wanted to go bigger, and they wanted to go better, and so a group of engineers at a number of different laboratories and defense contractors came up with the most formidable system of them all, Gorgon Stare.

A little note on the name. This book kind of sells itself because it has this name Gorgon Stare on the front. I didn't come up with that name. In fact, it was a little hard for me to find out who had come up with it. It seems like it was a retired Air Force officer who had the brilliant idea of naming this all-seeing camera after a Gorgon. Those of you who are familiar with Greek mythology will know that a Gorgon is one of these three mythological creatures, the most famous of which is Medusa, who has a hair of snakes and is able to turn her adversaries to stone merely by looking at them.

I came across this Pentagon manual for surveillance that said, "It is our intention with these technologies not just to watch our adversaries but to give them the sense that we know even their intent, so that they are always looking over their shoulder." As you can imagine, naming your technology after a Greek mythological character who can turn her adversaries to stone would be very effective in achieving that goal.

Gorgon Stare is a truly massive system. It is very much an evolution from those early versions, and it is able to record such a massive view thanks in large part to a technology that most of us have in our pockets—and, if we all listened to Joanne, hopefully turned off—which is the cellphone. Cellphones have these tiny, tiny cameras in them, and that's the case because we all want to have a camera in our pocket, and we don't want to carry around a huge device. Well, if you get enough of those tiny little cameras and stitch them together, you create a truly godlike view of the ground, and that was the key innovation that gave rise to this technology, Gorgon Stare.

Gorgon Stare and a number of systems like it are still in use today in foreign battlefields. We know that it is operating in Afghanistan, we know that it is operating in Syria, and there are no doubt a number of classified systems that are operating in other theaters. This is a technology that has very much entered the pantheon of American spycraft, and to be sure, it is making a big difference in the battlefield.

But that is not the end of the story because very soon after this technology was first developed those developing it looked home. They realized that the technology would have tremendous benefits for use domestically, not only in law enforcement but also in search and rescue and disaster response and environmental monitoring.

I'll give you one example. In the aftermath of Hurricane Katrina, thousands of residents of the city of New Orleans were stranded on their rooftops, and so the responding emergency agencies had to spend hundreds of hours with dozens of helicopters literally scanning the whole city to try to find these people. With a camera like this, you park it over the city for, say, an hour or two. If it has an infrared capability, then you'll be able to pick up every single person stranded on their rooftop with a single aircraft.

Of course, though, the big one is law enforcement, and there has been a concerted effort by a number of manufacturers over the years to deploy the technology for law enforcement within our shores.

A few years ago I was riding my bicycle home in Brooklyn at about 3:00 in the morning, and I witnessed a shooting. I saw four people shoot a young man right in the middle of an intersection. The assailants disappeared into the night, and as I learned later this was an unsolved crime. The police were never able to track down the assailants of this poor young 19-year-old, who fortunately did not perish.

It occurred to me that a technology like this could have solved that crime because just like an IED attack you don't need to know where it's going to happen ahead of time, you can see where it happened in the footage and then rewind and find where those four people went to. You can also rewind further and see where they came from, and with that you will have some really solid leads on which to base your investigation.

For that reason, it's not surprising that a number of cities have invested in this technology, and while we don't have one city at least publicly to date at the moment flying this technology as we speak, there have been some very extensive programs in the past, the most extensive of which happened in the city of Baltimore in 2016. That year a billionaire philanthropist from Texas by the name of John Arnold, who likes to fund experiments with cutting-edge technologies that could have transformative potentials for law enforcement, made a donation to the city of Baltimore so that they could run this aerial surveillance technology over the city.

Because it was a donation and did not come from the government budget they did not have to seek any approval from the state attorney, from the mayor, the city council, or the state legislature. This was a little bit like the rogue cell in the NSA that is depicted in Enemy of the State, a kind of "going it alone" operation.

I was lucky enough to actually visit that operation when it was still secretly surveilling the city of Baltimore in 2016, and I sat in the operations center as they have this commanding view of 32 square miles of the city. They tracked hundreds of crimes. They used some interesting loopholes, by the way. The city has this policy where if they have any surveillance footage that is not related to an ongoing investigation, they have to delete it within 30 days. Of course, if you record the entire city at once, there are no doubt many crimes that have happened on any given day, and so you can hold onto this footage indefinitely.

I sat in on a briefing where the analyst from this company showed detectives how they had tracked a group of murderers who had shot down a 31-year-old man in broad daylight away from the crime scene. I was not supposed to be there in that briefing, by the way. I sat very quietly in the back because those were the instructions given to me. At the end of that briefing, one of the detectives said that it was the best briefing on an investigation he had ever seen in his life. Another detective was so blown away he almost didn't have any words. He said, "Oh, my goodness! It's like that movie, Enemy of the State."

By the way, it was only with the publication of this book that it became public knowledge that the technology was indeed inspired by Enemy of the State. It has such an uncanny resemblance to the movie that it was always used as a way of explaining what the technology looks like. Isn't that funny? When I found out that it actually was for a reason that they are so similar, I pretty much fell off my chair.

Perhaps the strangest thing about being in Baltimore was stepping out onto the street after visiting this operation center on the second day having just a minute earlier watched the entire city, zooming in on neighborhoods, tracking random cars here and there. I stepped out onto the street. There wasn't a cloud in the sky. I tried seeing the aircraft, but it was operating so high, over 10,000 feet of altitude, that I couldn't see it. But I knew that it could see me because I had even zoomed in on the very street where I was then standing, and that felt uncomfortable enough. But much more uncomfortable was looking around me at all the other people who were being watched by a technology that they probably didn't even know existed, and they had no idea. That to me just felt wrong.

Incredibly, none of what the Baltimore Police Department did during that operation was illegal. In fact, if any of you wanted to get in an airplane with one of these cameras and surveil the people of New York, it would be within your rights to do so as long as you observe the airspace regulations because the airspace is public. If you point your camera out of the window when you're on a commercial flight from coast to coast and take a pretty picture of the landscape, what you've done is not illegal. You're exercising your First Amendment right to record images from public space.

That law, or rather lack of laws, was developed at a time when this technology did not exist. It was developed at a time when really in order to get into the sky you needed a great deal of money and resources, and the technologies that you used once you were up there would certainly not be able to record an entire city at once.

That is where things stand now. There is nothing from a legal perspective or a legislative perspective blocking cities around the country from adopting this technology.

Indeed, when the Baltimore operation was revealed, there was so much public outrage that it had been a secret, that it was canceled but the man behind that company—Persistent Surveillance Systems—has not given up and he actually just did a series of interviews last week where he said he now has his sights on St. Louis.

St. Louis is also relevant to our story because it brings up an anecdote that again speaks to the danger and power of this technology. I spent some days with another engineer who is developing this technology. He flies it in his private airplane. I went up in his airplane, and we needed no special permissions.

He told me a rather troubling story. He said that in 2014 he took his airplane up to St. Louis, where the Michael Brown protests were happening following the shooting of Michael Brown, a young black teenager. He said he just did a series of tests over the protests. Be that as it may, he had also told me in separate conversations that he vehemently opposes the Black Lives Matter movement. He had even referred to the protesters, who were largely protesting peacefully, as "thugs."

Surely, there have been times in history when protesters, who have every right to do what they were doing, were referred to by the authorities as thugs, and we should be very glad that those authorities did not in every case have access to technologies like this that could allow them to act on those beliefs. We have controls in place for this very reason.

Many people ask me how I was able to get access to all of these fellows who work in this secret world of clandestine aerial surveillance, and to be sure, when I embarked on this project no one believed that anyone was going to speak to me. But everyone did speak to me, and there's a very clear reason for that. It's because they all understand that this is an important story with implications for all of us. They knew from the very moment that they first conceived this technology that it was dangerous and that if it was going to be deployed in our skies we were going to need real measures to ensure that these kinds of abuses don't happen.

Why else is it dangerous? We remember the name Gorgon Stare. We remember its desired effect, to make sure that the enemy is constantly looking over their shoulder. Do we want to live like that? Do I want to be looking over my shoulder if I legally attend a protest and exercise my right to assemble? Do I want to be looking over my shoulder when I attend a place of worship? Do I want to be looking over my shoulder going about my business, doing something that I will find out only becomes controversial down the line, because if you have access to a person's history you have access to their deepest secrets.

To be sure, there are criminals out there, and if we have ways to stop those criminals it is incumbent upon us to do so. I would very much like to see the assailants of the man I saw shot in Brooklyn be brought to justice. But it has to be done under a strict set of conditions, and there is no one I spoke to, even the most vehement proponents of this technology, who would say that there shouldn't be some form of regulation to ensure that those abuses do not happen.

By just talking a little bit about where the technology is going, writing a book about something like this is sort of like aiming at a moving target that is three years in the future and is very, very small. You have to be able to get a sense of where things are going, and it is very possible that by the time you publish the book things that you said were going to happen in the future have already happened in the past.

There are a few key things that are happening. One is that the technology is getting much smaller. A lot of the people here will remember how big cellphones were not that long ago. Now we have this tremendous capacity in the palms of our hands. The original cameras weighed upward of a thousand pounds. Now you can get exactly that same capability in a 30-pound system. What does this mean? It means we can put them aboard small drones. I spoke to one engineer who said that he could assemble a swarm of drones that would be able to surveil the entire island of Manhattan with this wide-area view. Do you know how many drones he would need? Seventy-six. That's not a great number of drones. If each one costs a couple of thousand dollars, that is so much cheaper than a single one of these original cameras. Somewhat troublingly, I could go online on Amazon tonight and buy a drone myself.

Another thing that's happening is that the technology is being automated, and that's something that we really need to think about, too. As one source told me, "We all like to joke that we were creating Big Brother incarnate, but really it takes a million people to watch a million people. You record this whole view of Manhattan, and if you want to watch thousands of vehicles, you need people clicking on the little pixels on the screen, tracking people wherever they went." That is incredibly labor-intensive, obviously. What if, with artificial intelligence, you clicked on a vehicle and said, "Tell me everywhere it has been and everywhere it's going." Not only that, "Tell me every other vehicle that is associated with this vehicle." Now the work that took the CIA weeks in order to unravel this network can be accomplished automatically.

There's more, though. Say you want to prevent crimes before they happen. Well, there are a number of things that will tell you that a crime is about to occur. Vehicles that are associated with violent assaults, for example, are often seen to exhibit a very consistent set of driving behaviors in the lead-up to the crime. They will drive around aimlessly. They will do sudden U-turns in areas where people are generally not prone to exhibit that type of behavior.

What if you said to this computer vision algorithm, "Tell me every time a car exhibits suspicious driving behavior in this city," and then it will be able to tell you every time a person does a dodgy U-turn. In fact, when I was on my way to a company developing that very technology I got lost in my car, and I had to drive around aimlessly for a while and exhibit the very behaviors that would have flagged me as a threat.

So you have to keep in mind that while the technology that we already have is extremely formidable, once it is automated you really reach that holy grail of all-seeing surveillance. Fortunately, these systems have not been fully automated yet. In that regard, we are still somewhat ahead of the curve. We are also somewhat ahead of the curve in the sense that it hasn't been deployed widely in U.S. cities yet.

I know that seeing all these technologies emerge, particularly in the context of all the other surveillance technologies that exist—Facebook and Google gathering all our data, closed-circuit TV (CCTV) cameras everywhere, license plate readers—it's very easy to be pessimistic. In fact, I had a rather exhausting interview with someone who clearly had no hope for the future. They said, "We are headed toward a total surveillance state, and there's nothing we can do." If everyone agrees that we need regulation, my contention is that it is the people who are optimistic who will actually make that legislation happen because people who are pessimistic about our prospects have no reason to act.

On that note, I would just share the fact that even though I spent the better part of four years of my life investigating technologies that generally kept me up late at night with visions of an overwatched future, I remain very positive about our prospects—but only if we face this future collaboratively with a mutual understanding that we all perhaps have more in common than we might think and that there are solutions to these problems.

With that in mind, I'd be very glad to take any questions that you have.

Questions

QUESTION: Hi. My name is Bill Rabin [phonetic].

One, have the Chinese stolen this yet? And two, when a drone gets shot down, does the technology self-destruct?

ARTHUR HOLLAND MICHEL: These are two very excellent questions.

Of course, Enemy of the State is available for purchase in other countries, too, so it was inevitable that other countries would take a great interest in this technology. There are numerous scientists and engineers I found in China who are working on wide-area surveillance technologies. Some of them have interestingly collaborated with Air Force engineers in the United States. They seem to be particularly interested in the automation of these technologies, but I would be surprised if there wasn't some sort of capability like this in the People's Liberation Army (PLA) Air Force.

This technology is also emerging at the ground level, so we're starting to see CCTV cameras of the kind you would see in your corner deli employing this same principle of wide-area surveillance, where instead of having to rely on the camera being pointed at the right place at the right time, you're able to see 360 degrees at once. There are a number of Chinese surveillance camera manufacturers that actually have those very capabilities built into even some of their cheap models. There's one company, Hikvision, that is becoming very quickly a global leader in CCTV technology, and they have a camera that's $1,500 that gives you that 360-degree view.

The second question was about whether the technology self-destructs if the drone that is carrying it is shot down. We are all familiar with what happened a couple of days ago with Iran shooting down a U.S. reconnaissance drone. I should note that that drone as far as we know was not equipped with exactly the technology that I am talking about here. It is nevertheless equipped with some very powerful aerial surveillance technologies, and I would imagine that there would be some measures taken to ensure that at least the very innards of the technology cannot be replicated, but that would probably be on a case-by-case basis.

It's generally understood that if something is that sensitive it's made very difficult to replicate it if the wrong people get their hands on it. But again, this is all very mysterious, so it would be hard to know in the case of Iran where the drone came down, how much of it was destroyed as it was falling from the sky.

JOANNE MYERS: You started to talk a little about the man in St. Louis who was surveilling Ferguson. What was he planning to do with some of the photos he was taking? You indicated there was sort of a nefarious purpose to what he was doing.

ARTHUR HOLLAND MICHEL: He said that it was only an experiment, that he wanted to test some tracking algorithms in an environment where you have a lot of people operating simultaneously on the ground. But I asked him a question. I said, "Well, if you had seen something that concerned you, would you have passed that information on to the police?" And he said, "Of course. That is my responsibility as an upstanding citizen."

If you're the kind of person who calls peaceful protesters thugs, if you're the kind of person who vehemently opposes a protest movement, then when you fly over that protest you're probably going to have a different threshold for what counts as actionable intelligence. He said he did not pass any of that information over to law enforcement, but he was very much disposed to do so should the need arise.

QUESTION: John McAuliff from the Fund for Reconciliation and Development.

Two questions. One is other potential actors besides the Chinese—the Russians, the Brits are already thoroughly wired, thoroughly camera-ed in London. Are they moving in this direction?

The second question—and I missed the first ten minutes, and you may have made this clearer—is your sense that it's inevitable and therefore you come to the question of mechanisms of control, court authority, or something like that, or do you think it's still possible to somehow prevent the technology being installed?

ARTHUR HOLLAND MICHEL: We are seeing some growing international interest in the technology within the United Kingdom, for example, increasingly in Europe. Israel has a manufacturer that develops a very wide-area surveillance system that it is actively marketing both for military and civil use.

Another thing I've noticed is that the very same principles embodied by this technology are starting to crop up in other kinds of technologies. You can think of this technology as being a little bit like a big data collection tool. The thinking behind big data is that you collect everything with the understanding that most of it is probably not going to be all that relevant but that it is only by collecting everything that you will catch those—to quote Donald Rumsfeld—"unknown unknowns" that could become relevant down the line.

We're also seeing the same principles of this technology heading up into space where, now that the technology is small enough, now that satellite technology is accessible enough, and you can launch large constellations of satellites relatively inexpensively, you will now have satellites conducting video surveillance of extremely wide areas from space.

The somewhat depressing conclusion from all that, which goes to your second question, is that if you have enough satellites doing enough wide-area video surveillance from space, you get to a point where the entire Earth is watched consistently at all times.

QUESTIONER [Mr. McAuliff]: By more than one country at a time.

ARTHUR HOLLAND MICHEL: Yes, by more than one country. Think about satellite imagery, satellite photographs. We already have huge constellations that produce pictures of most of the world at least on a daily basis. If you just extrapolate that a little bit further, look down into the future a little bit further, once those satellites have cameras on them, once you have enough satellites that where one satellite sort of disappears over the horizon and there's another satellite coming back over the other side to pick up the slack, then what we get is Google Earth, but moving.

In fact, a lot of these manufacturers have said that they think a workable business model for what they're doing is to just release this to everybody and by everybody paying for it to do any number of things—fighting crime, doing insurance appraisals, tracking wildlife, monitoring the morning traffic—

QUESTIONER [Mr. McAuliff]: Divorce.

ARTHUR HOLLAND MICHEL: Divorce settlements, of course. The cost is spread among everybody.

There's this real sense that there are so many possibilities here that if you just put it out it's like the Internet. The creators of the Internet didn't know everything the Internet would be used for, but they had a sense that it had these boundless possibilities. In that regard, it does feel a little bit like the cat is out of the bag, so to speak.

A lot of these enabling technologies that make this surveillance possible come from other sectors. We have the phone camera that enabled this technology, but we also had the gaming console, which also played a fundamental role in the development of this technology. With these PlayStations and Xboxes, you have these little devices inside that are able to process vast amounts of imagery, and that was enabling for this surveillance, even though the creators of that technology didn't have surveillance in mind.

So, my sense is that, yes, we cannot stop the onset of technology. We need to keep up with it.

JOANNE MYERS: Doesn't this in some way impinge upon our Fourth Amendment about searches and that people could challenge that it's sort of an illegal search in a way, following people?

ARTHUR HOLLAND MICHEL: There was a Supreme Court case recently that touched on that very issue. It was the Carpenter case, which I encourage you to look up. The Court determined that the police using historical cellphone data records to see where a person had been over an extended period of time was a violation of their Fourth Amendment rights because they said that while a person's location at any one particular moment in time is not private—I am standing before you here at E. 64th St. in front of a bunch of people; that is not a private thing—knowing where I am at any given time over weeks or months is, of course, private. So they determined that accessing our location data in that way was an unreasonable search.

But they declined to adjudicate on whether that counted for live surveillance. So, if some government agency is tracking my cellphone right now in real time, they're able to do so without a warrant because the Carpenter case did not come down on live tracking.

They also didn't extend their ruling to other technologies that could achieve a similar outcome. My sense is that this technology that I talk about, an imagery-based version of the cellphone tracking, could be challenged in that way, and I think there would be a reasonable case for that challenge. But it still hasn't happened yet, so the jury is still out, if you'll excuse the pun.

QUESTION: Hi. I'm Saar Yaniv. Thank you so much for the presentation. It's quite an eye-opener.

You mentioned that this was used in Baltimore in 2016 and that there are plans on using it in St. Louis I think. When are we going to see it over New York City, Los Angeles, San Francisco, Moscow?

ARTHUR HOLLAND MICHEL: It's a tricky question. The Los Angeles County Sheriff's Office has experimented with the technology. I'm sure the New York Police Department (NYPD) is aware of it. There has also been a great deal of interest in the technology from federal agencies, the Department of Homeland Security, and the Federal Bureau of Investigation (FBI).

For this technology to emerge over a city like this, a few factors have to converge. You have to have as the industry would call it "political will," so you have to have the NYPD saying, "We're going to do this, even though it might not be that popular." You're going to have to have a strong sense that it will fit in with the other surveillance technologies that are at play. There is, of course, a cost factor that one needs to keep in mind. There has also got to be the fact that people will allow it to happen.

All of the programs that have happened to date that were publicly known were stopped due in large part to public pushback, which raises a very interesting possibility: Maybe we've suddenly reached a technology that is just one bridge too far, a technology that people say, "Actually no, we don't want that. We're okay with CCTV cameras. We're sort of okay with license plate readers. But we don't want an unblinking eye in the sky. Don't do that." I think that's a very real possibility. There has been a lot of action to take down traffic light cameras because those are seen as a violation of people's right to run traffic lights.

On the other hand, there is a sense that even if it is not this all-seeing eye in the sky that ultimately ends up over New York City there will be other versions of technologies that enact on this very same principle.

Let me give you an example. We're all familiar with the concept of face-recognition technology. If you have a dense-enough network of CCTV cameras—and certainly New York City has a very dense network of CCTV cameras—then a person who pops up on one camera can be identified by the police, and the system can say, "Recognize who this person is. Okay, it's Arthur Holland Michel." Then, the analyst says, "Tell me every time you recognize that face in every other camera that's in the network." Suddenly, you don't need an eye in the sky. The cameras that are already around us that we have accepted into our lives become this way of tracking us persistently through time and space.

That is why we need to also be conscious about the emergence of facial-recognition technology in law enforcement, something that has already happened. In fact, the NYPD commissioner recently had an op-ed in the Times about how facial recognition was a very positive thing for law enforcement, something that was difficult to abuse. There are not strong regulations for the use of facial-recognition technology, either, so that should likely be part of the conversation as well because if it's not this one thing that takes us to that future it'll be another.

QUESTION: Hi, Peter Burgess, TrueValueMetrics.

I'm extremely depressed by what you've been talking about, but I think there might be a silver lining, and that is that you're tending to focus on what I would describe as the individual misbehavior. But there's a bigger thing on this planet, I would argue, and that's the corporate misbehavior. There is a huge amount of pollution that emanates from the corporate world when they think we can't see them, and this is a massively powerful tool perhaps to stop the corporate world doing nasty things.

ARTHUR HOLLAND MICHEL: Oh, yes. Absolutely. That's an excellent point. I'm glad you bring it up because again we can't lose sight of the fact that these technologies do have some positive potentials.

One of the activities in Baltimore was actually to track illegal dumping, which is not necessarily corporate pollution, but it is a version of that. It is people abusing their privacy, if you will, to put things out into the environment that they shouldn't.

It also raises another point that I didn't get to, which is the idea of using the technology to conduct surveillance on the state. In Baltimore there were a number of police shootings that were recorded by the camera, and that raises a very interesting possibility. Now you can use the camera to see if the police's account of what happened on the ground is actually true. You can make sure that the police are not violating investigative principles.

The camera also was used to clear a person who had had their house searched because there was a warrant stating that it was probably involved in the drug trade, and that was based on some ground-based surveillance that said that there were people coming and going to this house all the time. If you look at the aerial footage, you only see a couple of people showing up at this house once or twice a week. So, then you can challenge the government on that count.

The question again is how we balance that against someone who uses that very same capability to, as you mentioned, perhaps spy on your spouse to make sure what they're doing is in line with your desires and your principles.

QUESTIONER [Mr. Burgess]: I'm a bachelor, so it's okay.

QUESTION: My name is Kevin McMullen.

I have a question as an outsider. In high school I took Latin instead of physics and chemistry, although I do have a push-button phone now and a VCR, but how can these systems be sabotaged by another government or a private organization? If the Taliban should knock on my door tomorrow and want some advice, what could they or a foreign country do?

ARTHUR HOLLAND MICHEL: Not to put any ideas in anyone's head, but there are a couple of things. One of the concerns about this technology is that it won't be compromised when it's airborne but rather that it will be compromised once the data is on the ground. So, as I mentioned earlier, Baltimore can hold onto this footage of the city indefinitely. Think about that for a second. If someone hacks into that, they're able to zoom in on their house and see where their wife went in the morning. They're able to track their political opponents, say.

Imagine a white supremacist watching the protesters in the aftermath of the death of Freddie Gray, which happened in Baltimore in 2015. They can track every single protester back to their home, and now this extremist organization has the home addresses of all these people who they know to support a cause that they oppose. We really need to think very carefully about how the data is protected.

In terms of how it's used in live operations, though, adversary groups have shown some capacity to intercept the data as it travels from the aircraft to the ground. A group of Iraqi insurgents in I think 2009, using some equipment that could be bought for $30 at Radio Shack, were able to actually intercept the video feed from a Predator drone. The United States quickly patched up that vulnerability, but this is information flying through the air. It becomes a cyber issue, and as we know cyber is a game of catch-up, and you can always have an adversary that gets one step ahead of you. I would say that if the Taliban were to knock on your door tomorrow you might note that as one place that people have looked to.

QUESTIONER [Mr. McMullen]: What about the authorities altering—I never saw the movie Forrest Gump, though I did see an Agatha Christie's Poirot in which he and Captain Hastings were inserted into a crowd near Queen Elizabeth. Can the authorities digitally change these recordings so as to implicate someone?

ARTHUR HOLLAND MICHEL: Oh, yes. Certainly.

The imagery is pretty basic once you get down to seeing individual people in cars. Because it's watching such a large area a human might only appear as a little dot, which is all you need, by the way, because you just want to figure out where they've gone to, and once you know where they are, then you use another camera to get a close look at them.

In theory, someone could add a couple of dots to the imagery, which is why there is a lot of talk about the importance of the custody of the data, making sure that it is always known who has access to it and when and for what reasons. It is also important to think about the custody of the data when one looks to prevent, say, a rogue police officer from just logging into the system one night when no one's watching and tracking someone, which is very much a concern.

The way they deal with that is that every time—some of the companies already do this, actually—someone logs into the system there is a record of when they logged in, who it is, what they looked at, how long they looked at it for, how many times they accessed it, so that if a person were to abuse it in that way there would be very much a paper trail that can be challenged.

Another principle that is somewhat related to this is this idea that people have a right to access surveillance data that is collected on them, and so that can be very interesting in this case. You say, "Well, I want to access this video that you say you have of me traveling to this place," and then you can challenge that. That would be another way to add a measure of accountability to challenge some of these things that you're talking about.

JOANNE MYERS: I think you've encouraged all of us to be very good every day, all of the time, everywhere we go.

QUESTION: I'm David Hunt.

Several years ago, I had occasion to go to San Diego on a Hudson Institute program, and among other things we visited a company called General Atomics. General Atomics is the company that as we understood it is the lead producer of the drones—the Reapers, the Predators—these incredible pieces of equipment that can hover at 30,000 feet for 24 hours.

Is what you're really talking about this equipment being used by civilian authority in a more miniaturized way? That's really what it comes down to, I think.

ARTHUR HOLLAND MICHEL: Sure, sure. It's a good question.

When we talk about drones, you have to sort of bifurcate your thinking a little bit. We have the military drones like the Predators and the Reapers that are very, very large, in some cases as large as a commercial airliner. Then we have, on the other end of the scale, the type of drones that you or I might buy on Amazon for $1,000. These drones are very, very small, they're very, very light, and, of course, as a result they can't fly for nearly as long, they cannot fly at the same altitude, and they cannot carry large powerful cameras.

At the moment, law enforcement does not have access to those large military drones. In fact, you cannot use those large military drones in U.S. airspace unless you're the military, and you have to have some pretty exceptional permissions to do so. For that reason, law enforcement generally uses this technology aboard manned aircraft, piloted aircraft.

But there is going to be a point where that line is blurred very much because as the technology becomes smaller and as the skies open up to larger drones—which is happening; this fire season in California you have no doubt that CAL FIRE is going to make extensive use of military Reaper drones—then we will see drones operating in our airspace for persistent surveillance.

There are already at this point over 900 public safety agencies in the country, most of which are law enforcement agencies, that already operate those very small drones. That feels very much like a sort of stepping stone to a larger capability, but then again, as I said, if the technology is now 30 pounds, we may get to a point in a couple of years where that 30-pound technology can fit on the very drone that you or I could buy on Amazon.

QUESTIONER [Mr. Yaniv]: I think you mentioned something about California. Did you say something about the wildfires there? Is that what you mentioned?

ARTHUR HOLLAND MICHEL: Yes, yes.

QUESTIONER [Mr. Yaniv]: So I assume they're using that in real time to detect fires and dispatch firefighters to those locations right away. Are there any other instances besides wildfires that any other local or federal authorities are using in real time?

ARTHUR HOLLAND MICHEL: In real time? I'd make a slight adjustment on what you said, which is that these drones are used once the fires are known, and they are dispatched to keep an eye over it and to help coordinate the movement of the firefighters on the ground.

QUESTIONER [Mr. Yaniv]: They're not used to detect fires.

ARTHUR HOLLAND MICHEL: They are not, no, no, no, because you would need many drones over an extremely wide area.

The Forest Service has acquired a small number of these wide-area cameras, which is really interesting, because then you can watch the whole fire at once, and you can see how it moves. It was very interesting when I talked to the Forest Service. They talked about the fire as an adversary, and they used many of the terms that the military had used in describing human adversaries. They talked about understanding the "pattern of life" of the fire, for example. That's going to be another big one, I think.

There was another application—perhaps to end this on a positive note—of a company using a wide-area surveillance technology to find polar bear dens on the North Slope in Alaska because when it is drilling season they want to make sure that these polar bears are not disturbed. So, if you can identify the locations of each of these dens, then you can make sure that they get a full season's sleep.

Maybe we should think about that as we try to fall asleep tonight ourselves.

JOANNE MYERS: Thank you very much for taking us into the 21st century. Thank you.

ARTHUR HOLLAND MICHEL: Thank you so much. I really appreciate it.