New Frontiers in Intelligence: The Changing Nature of Espionage

Panelists discuss the new frontiers of intelligence gathering, examining how emerging technologies are transforming espionage, expanding its reach to the public domain, and reshaping the future of national security.

GELLMAN: Welcome, everyone. I’m Bart Gellman. And I’m here with Robert Cardillo, remotely, Anthony Vinci, and Rebecca Weiner. You’ve got their bios. I’ll just say, we are very lucky to have—I mean, these are the elite thinkers and operators of the intelligence community. And we’re here to talk about “New Frontiers in Intelligence: The Changing Nature of Espionage.”

This meeting is on the record. And we’re joined—welcome here to the seventy-or-so members in the room, and I’m told we have about 400 or more tuning in virtually. We’ll take questions from members in both groups halfway through this event, or thereabouts.

I want to start with Anthony, who’s written a really outstanding new book, which I have not yet finished but I am very impressed by. It’s called The Fourth Intelligence Revolution. And I wanted to start by asking you what that means.

VINCI: Thanks. And thanks so much for having me. And thank you all for joining today.

I look at intelligence from a historical point of view in the book, to start. And thought about it as what has changed over time and how has there been revolutions that fundamentally change what intelligence does, who’s targeted by intelligence, how the process works, who’s involved in it. And I saw the World War II really sort of the big bang of intelligence, where we had intelligence in America, obviously, before that. Even George Washington was running spies during the Revolutionary War. But in World War II we created what most of us would consider to be a modern-day intelligence agency, in the OSS. And that set us on a path that we still see the ramifications of. That it’s a very similar culture and structure in intelligence agencies.

Then there was this sort of second change, this revolution in intelligence after World War II, with the Cold War. Where America had to professionalize and really expand intelligence. We created the CIA. We created the NSA, DIA, and other agencies. And that worked for several decades until 9/11. And 9/11 forced us to create a new form of intelligence because that old one, the way we were doing intelligence during the Cold War, was so good at what it did, we compartmented so well and were so secretive, that it actually created cracks through which al-Qaida was able to target America, because the CIA and the FBI wouldn’t talk to each other, and other agencies want to talk to each other. And this new threat in these networked terrorist groups were able to prey upon that.

And so we had to change the intelligence world again, and expand intelligence into this whole-of-government approach. And we created the Office of the Director of National Intelligence and the intelligence community that we see today. And that revolution really is where I came into the picture. I began serving in intelligence after 9/11. And went to Iraq, like many people did. And I think a lot of us considered that the be the norm in intelligence. And now what I’m seeing is that intelligence is again going through a revolution. And this time it’s driven by China as a threat, and by new technology, in particular by AI. And the book is really about what that revolution looks at—looks like, and how intelligence will come to look over the next several years.

GELLMAN: Thank you. Rebecca, we’re starting to talk about the collection end of intelligence. And I wonder if you could talk some about the evolution of the threat that we’re monitoring.

WEINER: Sure. First of all, thanks for having us. It’s interesting, picking up on your revolution, you know, before 9/11 NYPD, which is where I work, wouldn’t have thought of itself as having a real intelligence collection apparatus. That changed after 9/11. And convincing detectives, investigators, civilian analysts that work for a municipal law enforcement agency to have an intelligence mindset, connecting dots, building out networks, understanding the domain in a different way, on the one hand felt like a philosophical shift, but on the other hand very much in keeping with what our officers do every day. Nobody knows our streets better than the officers who patrol them, a department of 50,000 people that reflects the diversity of New York City. So it was an easy and very natural evolution. And one that has been replicated among law enforcement agencies around the world.

Which has been convenient, because the threat environment has also shifted radically, driven by some of the same technological advancements that Anthony covers in the book. Especially the last decade, our threats have changed radically. There is so much more to collect. This is a world in which the open sources are tremendously fertile for anyone who’s looking to put together the piece of a puzzle. When we have an incident that happens, within minutes or hours we are able to get 80 percent of the way there in understanding what has happened based on primarily commercial, open-source information. And then, of course, there’s proprietary information, sensitive information along with it. But we are wading in a sea of digital detritus that gives us a lot more noise, and also more signal, quite quickly.

GELLMAN: So hold that thought, because I want to come back to open source. Is the threat changing? I mean, we had—we had a migration from a near-exclusive focus on nation-states for a long time, and then to nonstate actors and lone wolves. Is there a further evolution since then?

WEINER: Totally a further evolution since then. And, you know, we think about counterterrorism specifically, when I started in 2006 we were focusing primarily on al-Qaida. In 2010 and ’11, that changes to a more homegrown, lone actor based model. In 2016 you’ve got tremendous ideological diversification. So you’re looking at foreign terrorist organizations and those who they inspire, but also political violence, the far left, the far right, conspiracy theory-driven violence. And over the past year we’ve seen an acceleration in contagion of grievance-fueled violence. You know, we had December 4 in this city the assassination of a CEO of a health insurance company, allegedly, by Luigi Mangione. That spawned a torrent of threats in the online world that amplify the murder, that encourage more such violence, and that results in more violence.

And, you know, a spate of political violence and assassination, some motivated by the war in Gaza. Assassination of Israeli embassy staffers in D.C., Charlie Kirk, Minnesota lawmakers, an increase in targeting of government employees, especially around CBP, ICE, and the issue of migration. Grievance-fueled violence, not just Luigi Mangione, a mass shooting at 345 Park Avenue over the summer, a shooting in CDC headquarters, a rise in nihilistic violent extremism, which is a relatively new category, which is carried out predominantly—unfortunately targeting and perpetrated by minors, often resulting in school shootings. We had one example in Minneapolis, an example in Wisconsin, an example in Tennessee. So these incidents of terroristic violence are becoming more fast, and much less predictable. And they are really happening against a backdrop of an online ecosystem that’s amplifying and encouraging more violence. And that’s where we spend our day.

GELLMAN: Let me—thank you. Let me move to Robert, who is joining us online. You were the first director—sorry, not the first—you were—you were director of the youngest, I think, or certainly a very young intelligence agency, the National Geospatial Intelligence Agency. And I just have to start with an elementary question. Can you tell us a little bit about what is geospatial intelligence? And what would it tell us, say, about, you know, the Upper East Side, or 58 East 68^th Street?

CARDILLO: Thanks, Bart. I’m happy to. And happy to join my colleagues on this panel for what has already been an interesting discussion about what I consider to be an essential book. Thank you, Anthony, for writing it.

So to your question, Bart. Geospatial information—and I say information before intelligence—but geospatial information is the way that we become advantaged whenever we have more confidence or certainty about our location, and the locations of things that we’re either interested in or we could be concerned about. And so—and in my history that has been using space predominantly for that. And we should move beyond that quickly, but that was my early experience. Those space systems, at the time, were almost all government owned and government operated. And almost all were held at the top-secret level. So the way I used to say it is, if you had a top-secret clearance you could ask me a top-secret question. I could go to a top-secret computer, to ask a top-secret satellite, get a top-secret image, exploit said image, and send you back—just you—a top-secret answer. It was a very closed system, as you can imagine. (Laughs.)

But, to your question, that person wanted to know more, in those days, about the Soviet capabilities. You know, their Army, Navy, Air Force missiles. And we would use that technology in order to provide what I consider to be the essence of intelligence, which is advantage at point of decision. And we were able to provide that information to our decision makers and our war fighters as we experienced the Cold War to make better decisions. And so, back to your question on the Upper East Side, geospatial information at a much more, you know, everyday level can help you make better decisions about how you’re going to get from one place to another, or how you’re going to avoid an incident or traffic, and how you’re going to find something that you desire—could be a cup of coffee. And so as every—as important as location is in all our lives, geospatial intelligence is the very high end extraction and manipulation and presentation of many, many collection sources, again, to get people to better decisions, safer locations, better understanding.

GELLMAN: So when you took charge of the NGA it was, I think, less than twenty years old. And yet you believed that it had grown set in its ways and needed disruption. Which is why you hired its first ever chief technology officer, who’s on the stage with us. So why? What needed disrupting?

CARDILLO: Yeah. So, you know, I told you about kind of the mantra about—and the closed system, everything being top secret. I joined the guild, and I use that word specifically, of, you know, imagery analysis in 1983, in Ronald Reagan’s first term. I became the director in 2014 in President Obama’s second term. You know, we’ve just covered the dramatically changing landscape of the threat. What I was worried about as I stepped into the director’s chair is that, even though we were intellectually aware of commercial sources and foreign partners and, you know, video and unmanned aerial systems, I was worried about what I call the mental legacy, and the mental weight of what you did yesterday. And that resulted in a pride in the agency in which I shared, but I was concerned that that pride would inhibit us from questioning how we would succeed today, much less tomorrow.

And so you’re right, I was looking to disrupt what I thought was a comfort that was a bit false and somewhat dangerous. And through a friend of a friend, I found Anthony. I first invited him into my advisory board, an external board that does what it sounds like. But from about the first thirty-four seconds on, when I heard Anthony show up, I said, oh, this is different, because everybody else at that table were former directors, right, and former leaders. And Anthony, as he just explained in his experience, had come to this from a different point of view. And I knew immediately that that he was the kind of intellect, but more importantly he was the kind of motivational challenge that my team needed. And we can talk more about it, but I’m more convinced today, now that it’s eleven years later, that the seeds that Anthony planted not only changed the agency. I think they’ve—I know they’ve changed the intelligence community. And look, again, thanks for writing in the book, because I think—I’m hopeful that will compel another discussion about how much more we need to do.

GELLMAN: Anthony, I once had a conversation with Bob Mueller when he was—it was an interview with Bob Mueller—when he was FBI director. And we were talking about how an intelligence agency can hire the technological talent that it needs when it’s competing with very high salaries and other benefits in Silicon Valley. And he said one factor was that you tell people, you could do really cool stuff here that will get you put in prison if you try to do it out there. Is that it? Or, I mean, how did you attract the kind of tech talent you needed to an agency hardly anyone had heard of?

VINCI: You know, it is funny, because I started out as a case officer, which is someone who spots, assesses, and recruits sources. And you are—I would describe myself as a paid criminal, you know, by the United States government. Because they would say, go to this other country, break all of their laws, essentially, get their most trusted people to commit treason on them, and bring us back the information. So it’s true what you say. And you do do very cool things. And you get to work with, on the technological level, data and tools that just aren’t available to the rest of the world. So, you know, to use geospatial, there are amazing commercial satellites out there right now, and companies using these satellites. But when you come into the fold, into that top-secret world Robert was talking about, there’s even more amazing stuff.

And so when we wanted to recruit people—and at that time NGA, for example, had less than ten data scientists, full-time government employee data scientists. And I proposed a plan to Robert where we wanted to get eighty-five. Eighty-five data scientists and engineers and people working on data and AI. And it was considered insane because not only do we have less than ten, but the turnover rate was, like, 25 percent a year. So, like, we kept losing them too, for the reason you brought up. Which is, people get paid a lot more money to do technology in the private sector. But I think that when you come into intelligence the reason you work there is for a couple of reasons. One is out of a sense of patriotism and desire to help your community and your nation. And that’s a baseline. You’ve got to want to do that at some level. Many people do. Actually, many people do in Silicon Valley, and did then.

But when you talk to engineers about their—literally their job, engineers aren’t always just in it for the money. They’re in it to build amazing things and build things nobody else has ever built before. And that really was the thing that we sold to people when they came on, including our first chief data scientist who was—did an outstanding job for the agency. He had a Ph.D. in data science, which was rare. He might have been one of the few people in any agency at the time who had that. And that was really what sold him, and many of the people who ended up working for him, was, you know, you’re going to do—these are going to—this is going to be amazing data. It’s going to be amazing engineering problems. And you’re going to be solving things that are just cool.

You’re not going to be building advertising technology for, you know, a company that wants to push more ads on a social media site. You’re going to be doing something that, like, saves human beings from terrorists, or, you know, is able—you’re trying to figure out how to see underwater to see where submarines are, or how to spot satellites in space that—you know, before they come and attack somebody else’s satellite. Like, science fiction kind of stuff. And now I’m in pitch mode. So if anybody’s listening who’s thinking about—who’s an engineer and wants to join the intelligence community, reach out and we can—we can try to hook you up with a job.

GELLMAN: Rebecca, I want to come back to what you were saying about open-source intelligence. What’s new, or at least significantly evolved, about open-source intelligence? What can you learn from open sources now that you used to need a spy agency to find out? And what impact does it have that you’re not the only one learning it?

WEINER: I mean, it’s all about the open source now. And I think it’s hard to underline that point too much. The amount of information that is immediately available, not just by intelligence community, not just by law enforcement community, but by everybody in real time, is a source of strength for those who are trying to put pieces together and understand a story. But it also can be incredibly complex and dangerous. Intelligence has been democratized within the law enforcement, intel world, much more broadly. Redounded down to municipal law enforcement, as I was describing, but also within the private sector, also within the media. And as quickly as we’re able to put together the pieces, the water’s muddy. So information is very often not reliable anymore.

On election day we were dealing with a spate of hoax threats across the Northeast, in New Jersey and New York. Very sophisticated-looking fake ISIS propaganda that was accompanied by a threat. And the goal, of course, is to instill some fear and some chaos into polling sites. We were able to look at it and say, well, no ISIS magazine would have al-Qaida propaganda in it. This is obviously fake. But it was very quick to engineer. So it has become a shield and a sword. It’s useful for those of us who are gathering, but it requires human beings to discern both what’s real and what’s fake, and the intent behind it. Anybody who is in the intel world and not doing open source is missing two-thirds of the picture.

GELLMAN: Two-thirds?

WEINER: At least, I would say. I mean, it depends on your job, of course. If you’re working in geospatial, right, there are still certain elements of the intel world that are proprietary, sensitive, and altogether classified. But where we are, in the world of counterterrorism and in the world of manifested threats day to day, it is all about the open-source information and the smart humans who are able to make sense of it. And that point is really important too, because there’s so much discussion of AI and efficiencies in technology, and the technological problems that are going to be solved by technology. And in our experience it is really the human beings who are using the technology that are integral, whether it’s on the collection side, human intelligence collection, or on the analytic side, making sure that we’re getting not just fast information, but right information.

GELLMAN: What’s an example of a nongovernment entity that’s doing really good, high-quality open-source analysis?

WEINER: There’s a ton. Syrian chemical weapons is a good example, right? Some great work being done on that several years ago out of Bellingcat. And we talked about this a little bit on our call. Increasingly, private sector companies are hiring and creating in-house intel services which will flag threats to us on a daily basis. And they’re not just sifting through social media and sending us threat information, but they’re often doing what used to be the purview of mostly CIA, running sources. We have a lot of—and, again, information is collected. It’s being democratized. That creates some obvious problems you could imagine around coordination, if you have an undercover type of a source being run out of a private company or an NGO. So the world is totally different than it was even twenty years ago when I came into the intel discipline.

GELLMAN: Robert—

CARDILLO: Bart, can I add—

GELLMAN: Yes, please, go ahead.

CARDILLO: On this question, first, two-thirds is too low for geospatial, by the way. It’s at least 80 percent, if not higher, that is in the open and accessible. And I think growing higher every day. And the second piece, I would say to the end of your question, Bart, about what’s—what does it mean that the adversary or the threat also lives in that, you know, open-source milieu? I think that’s the essence of the challenge with the intelligence community, because, as I explained, you know, when I grew up we owned everything. So, hey, I have a satellite they don’t have. I have an image they don’t have, right? So advantage was in the—was in the source.

Now we’re all kind of existing in this, you know, overwhelming amount of source. And so you have to get advantage a different way than you used to. You have to operate differently. You have to—you have to analyze. You have to create the right algorithm. You have to create the right processes. And this is what—this is where—and, again, back to Anthony’s book—this is where we have to focus our attention and, vice, you know, trying to kind of fight the wave, you know? I wish it could go backwards, or I wish we could go to the old days. Those kinds of things are where I think gets us in trouble.

GELLMAN: So I was about to ask you, Robert, what you make of, I guess you could call it the open sourcing of classified intelligence? The phenomenon that’s grown up in recent years of mass leaks. And I—just to declare my position here—I was one of the three reporters with the Snowden files. And you were in the Oval Office, I gather, when Snowden first became public. What’s the cause and impact of this phenomenon of mass leaks?

CARDILLO: Boy, this is—we don’t have enough time to fully delve into this. Let me just say that, one, I’m not the first person to say that I think the U.S. government has a propensity, if not a disease, to continue to over-classify, at all levels. And, again, I think we do it for the right intent, but for the wrong reason. And it’s that mindset that I talked about that, oh, we’ve got a—we’ve got a special thing that no one else has. And so let’s put eight wrappers around it, and three code words, and all these things. And by the way, there are those things. Anthony mentioned a few of them. And we should do that. But they’re quite exquisite. And they’re quite, you know, I would say, small relative to the overall volume.

But, Bart, back to your point, I think leaks—you know, you mentioned Snowden, you know, WikiLeaks. You know, the Air Force, National Guard, you know, we have them cyclically. Look, I think, you know, Bart, you and I could have a separate discussion about, you know, did he do it the right way, you know, was there really a whistleblower that he could have turned to? And I know there’s a debate about that. But what I think we should take away from those is the fact that it is part of our ecosystem, this tension between, you know, open and classified, and protected and unprotected. And I think the U.S. intelligence community and our allies, we need to just get better at engaging in that—in that debate and that discussion.

I think when we—you know, when we throw up the, it’s classified we can’t talk about, I think it does a disservice not just to this specific debate, but to the broader debate. That if the public, if the taxpayer doesn’t have enough appreciation for why they’re spending—and I forget what the last year number was but it’s getting close to $100 billion a year in the U.S.—on the U.S. intelligence community, then, you know, you risk your—(laughs)—you know, your funding going forward. So I think we just have to get better at engaging in the debate. And I think, in some ways, leaks, again, while we can disagree on the instances of them, broadly end up supporting that debate, frankly.

GELLMAN: Anthony, I’m going to ask my last question before we move to member questions. And I guess I’m going to dump on you an uncomfortable one. I want to talk about the threat of politicization of intelligence. It’s not brand new. I mean, you know, Bill Clinton in 1993 attacked a national intelligence estimate he didn’t like on Haiti. Dick Cheney tried to change analytic conclusions about WMD. Is there something happening now that’s new or worse?

VINCI: Pass. (Laughter.)

GELLMAN: Oh, come on. There’s no pass.

VINCI: I think that you’re correct. There’s been politicalization of intelligence as far back as intelligence goes. And you can look at LBJ. You can look at JFK. You can—and you can even, you know, probably look back at FDR, and even World War II. So there’s always been this temptation. And there’s been a fear of politicalization of intelligence in America as well. There’s always this—been this innate skepticism of intelligence. It’s the reason we didn’t have a peacetime intelligence agency in America until 1947, because we didn’t—we don’t trust intelligence officers. And we still probably don’t trust them completely. And one of the reasons is we fear that a political party will use intelligence for political reasons. So I think we’re sort of right to fear—to fear it.

And it’s—now, so I don’t think there’s anything different at a fundamental level today. One can argue whether there’s more and whether this administration or the last administration, or the administration before that, or before that was politicizing more, or more overtly politicizing intelligence. Part of what may be going on is that more intelligence has made it into the public domain. We’re more willing to put it out there. But I think the thing that’s actually driving it is that intelligence is affecting the lives of everyday people more than it used to. And intelligence used to be something that was sort of reserved primarily for military and political purposes. It was about a Cold War with the Soviets. It was about wars and so forth. And now we’re seeing intelligence creeping into all of our lives because intelligence is being collected on Americans by Chinese or Russian hackers or intelligence agencies.

Information operations are being targeted on Americans and used by adversaries for trying to skew elections, like in 2016. And so we’re all part of this sort of great game of intelligence now. And we feel it and see it. And so politicians, who are there as the decision makers and dealing with it, are part of that too. And I think that’s creating this cycle where there’s more intelligence out there. We’re seeing more of it because it is more easily declassified and so forth, or leaked. And it’s affecting us more. And politicians are using it more. And that’s creating this sort of cycle, which is dangerous.

GELLMAN: OK, Anthony, pardon my interruption, but I want to get to member questions. I’ll start here in the room. Please keep your question concise. Keep it a question. And perhaps identify yourself before you ask. Go ahead, sir. Wait for the mic please.

Q: Hi. I’m Aleksandar Overholt. I’m with Victory Six Advisors.

The question that I’ve got is that, you know, we spend a lot of time talking about classified information and advanced methods for acquiring data. But I think that one of the biggest risks that we have right now is an ignorant population when it comes to giving up permissions on your phone when you’re installing an app and oversharing. So I think there’s two questions here. Is, how do we put some controls on the platforms or the data brokers that give basically open access not just to retailers but to our adversaries? And then the second is, how do we go back to bringing back the posters from World War II, which were loose lips sink ships? How do we educate the population?

GELLMAN: I once got a gift from a from a top defense official. It was a World War II soldier on a mug saying, how about a nice big cup of shut the fuck up? (Laughter.) So who wants to address this?

VINCI: Can I—if you don’t mind if I can hit this one, because I think it relates to the last question. I think that everyday Americans are now becoming a target of foreign adversary intelligence collection. And I think when you see healthcare data or financial data being stolen, or where you see information operations targeted at everyday people, that is now—this fourth intelligence revolution I refer to, I think that’s a big part of what’s going on. And the problem is that you don’t necessarily want intelligence agencies directly involved with the defense against those issues all the time, for fear of censorship and things like this. No one wants the CIA telling you what to read or not read on social media, or getting involved in your phone and saying, hey, install this app, it will protect your phone. Nobody wants that.

And so, to answer your question, I think that the only way to defend against it is to make Americans and people under threat like this—and maybe allies and so forth—more resilient by giving training, in the way that we give cyber security training, and maybe even to grade schoolers—which, by the way, we train grade schoolers on cyber security too—to recognize threats from data collection by foreign adversaries or information operations, so that they can start to think a little bit like intelligence officers, in a sense, to better protect themselves. And then install tools like that, because now they’re thinking about this as a threat, whereas I think for most Americans today don’t really think about this particular threat vector unless they live in the Beltway, or in D.C., or something like that.

GELLMAN: Thank you, Anthony. Oh, go ahead.

WEINER: I think—no, I think that there’s an important element, which is that, you know, there’s so much concern around government overreach and surveillance, but people are willingly giving private companies access to every bit of sensitive information about them imaginable.

GELLMAN: And that’s a national security threat as well as a privacy threat.

WEINER: It is 100 percent a national security threat, as well as a privacy threat. And the concept of ubiquitous technical surveillance is absolutely real. So people need to know that our government is, of course, constrained by the U.S. Constitution. And private companies, even in our own country, are not. So a first order of importance is reminding people—and the conversation around TikTok was really illustrative. People are much more concerned about being able to upload their videos and view them than the notion that a nation-state actor or others is able to have, as a giant treasure trove of information, sensitive information about the people holding these devices that we all carry around with us all the time. So, first, remember, we’ve already sold our secrets to the private sector before we worry even about nation-states.

GELLMAN: So I need to move to a question from our remote members. You can queue one up.

OPERATOR: We’ll take our next question from Prashant Yadav.

Q: Hi, Anthony. I’m Prashant Yadav. I work here at the Council. Congratulations on the book.

I am only halfway through so maybe this part comes in the second half, but I think we all acknowledge that the nature of biological threats is becoming more diverse and distributed. I would love your thoughts on how do we better coordinate OSINT and what private companies are doing around understanding biological threats, with coordinated action which the more standard intelligence apparatus is involved in? They cannot—we can’t afford them to happen as two separate activities. So how does this get coordinated, in particular for biological threats?

VINCI: This is the hardest one. To me, bio threats are probably the scariest possible threat against us. And only, you know, like COVID, I think, was just, like, a warning of what could possibly go wrong. So it’s absolutely something that has to be taken care of. But it’s also kind of the stickiest, most problematic area for defending against, because you’re talking about people’s personal health information. This isn’t just like PII, like, your name and social security number. This is, like, your medical information or even your genetic code. And so how do you—how do you track that and deal with that? And I think that the solution has to involve some form of public-private partnership. Intelligence agencies can support, you know, against state and nonstate threats building bioweapons, for example. We’re accustomed to that.

But if you want to get into bio threat monitoring and the spread of, say, genetically engineered disease or something, I think we don’t necessarily want, nor could we really have, a government performing all of those intelligence-type functions. And so you need to involve the private sector, and maybe even noncommercial private sector like nonprofits and so forth. And then that gets complicated because those types of organizations aren’t necessarily connected with intelligence agencies. So you have to deal with sharing information, and how to deal with the sharing of information. So this is a very complex issue, but to me the biggest possible threat, and one that we should be addressing immediately.

GELLMAN: I’ll come back to the room now. And the woman in white, please.

Q: Hi. Thank you so much. Kristen Kaufman from the U.S. Council for International Business.

Question for Deputy Commissioner Weiner. We have a new mayor coming in. How could the effectiveness of your group be impacted by this new mayor?

WEINER: I think what we’ve seen over the last twenty years, since I’ve been around, is that nobody wants an act of violence in New York City. Nobody wants an act of terrorism in New York City. There is a very active threat environment. I talked about one element of it. I didn’t talk about the conventional ISIS or Iranian lethal plotting, or all of the parade of horribles that we deal with on a daily basis. So I am confident that we will maintain our posture of doing everything that we can to protect our city, regardless of politics, and with the full support of our police department and our partners.

GELLMAN: Back online, please.

OPERATOR: We’ll take the next question from Hoyt Webb. Mr. Webb, please go ahead. Apologies. Seems we’re having some issues there. We’ll take our next question from Munish Walther-Puri.

Q: Thank you for great conversation. Munish Walther-Puri, principal at TPO Group, former director of cyber risk for city of New York. Rebecca, nice to see you.

My question is about, it’s 2025, talk to you about AI. How are you training the next generation of analysts to continue to use their intuition when they have a tool that can automate certain aspects of that. You know, as J.C.R. Licklider elucidated, computers are good at volume, scale, and computation. Humans are good at creativity and intuition. And so I would love to hear about how we put those together in an effective way. Thank you.

WEINER: Thank you for the question. You know, the story of the intelligence space is one of how humans are engaging with technology, whatever the technological advancement is of the moment. And so now it’s going to be AI. You know, we are still very much immersed in the sweat equity of smart people looking at facts and making sure that we’re choosing the right ones, and they’re taking the right conclusion from them. AI, hopefully soon, will be a force multiplier. The volume, as we’ve just been talking about, is so large, the amount of dust that we’re sifting through is so vast, that you need an efficiency generator of a computer to help you. You need something that can help surface the leads from everything that we’re seeing.

So I am hopeful that there will be an AI solution to the problem of volume and velocity that makes our work very challenging. I am doubtful, at this moment in time, that within the next at least year or two, or maybe five, you will have anything that would begin to displace the analytic judgments that are so central to our work, or the human intelligence work that is fundamental to discerning somebody’s intent to carry out a bad act. Because, of course, all of our efforts are on preventing the next thing from happening.

GELLMAN: Robert, would you like to take a crack at this, human intuition versus algorithm?

CARDILLO: Yeah. Yeah, we—you know, I struggled with it. In some ways, I think people—my teammates at NGA saw Anthony coming, and they pictured him with this dark cloak and this scythe, because he was going to—he was going to eliminate all the humans in the agency, right? We were just going to automate everything, and we would go full-on, you know, AI. And we were very purposeful. And I tried to communicate, let’s let these machines do what they’re really good at—at volume, at scale—answering three questions, what, where, and when, all right? And those are very important questions because it gives you the basics to find out what—or, at least to then elevate—and this is where the human comes in—the two more difficult questions, why and what’s next. And so it was—it was that kind of combination that we were trying to achieve. You know, let the machine do what it’s really good at, right? Now, let’s check it, let’s verify it, let’s grade it, and all those good things.

And then the other thing I’d say is that the—you know, there’s a way in which we need to figure out how to not pick one or the other, but team them. I mean, if you go all the way back to Big Blue (sic; Deep Blue), right, and Kasparov, right, you know, can a computer beat the world chess master? Turns out, the answer became yes. But whenever you took a chess master and teamed him or her, OK, with an advanced, you know, capability, they were better than everything—anything else. And so, you know, just as I think we heard, you know, from New York’s point of view, it really is not either/or, but how do we set—how do we set that teaming arrangement?

VINCI: I actually might be wearing the cloak, because I did used to think that, I agree. But the technology has moved so fast. I do see it as being creative now and having intuition in the near term, at some level. I think even—all of us I’ve seen where, by now, where you’re on ChatGPT, and you ask it to do something, and it does give you a new idea sometimes that you didn’t think about. And tools are going to be created for intelligence that do essentially all of the aspects of intelligence, including that. And the role of a human in the intelligence cycle, in particular for analysts, is radically going to change. And I think the person asking the question asked the right question, which is how do we keep our intuition sharp when these machines are able to do it?

And I would look back at, like, calculators as a good example. We used to have to do long division because we didn’t have calculators. Then we have calculators. And now we teach children long division just kind of learn the basics of math, not because they ever have to do it, so that they can become good at math at the higher level, at, like, theorem-building and so forth. We may have to do the same thing in intelligence, and teach people how to do intelligence analysis not because they’re going to be better at most of those aspects than the AI, but because we need there to be intelligence analysts who are able to do things like partner with a president and give them the why and to explain what’s going on, and so forth. And we need to still produce those people even when the machine is very, very good at this stuff.

GELLMAN: Back of the room, please. Gentlemen in the red tie, or reddish, maybe. There’s a mic—there’s a mic right there.

Q: Oh, sorry. I’m Kevin Sheehan, the managing partner for Multiplier Capital, a private investment fund.

Two of the themes from the panel are the increased reliance on a classification and increased reliance on open source. What does this mean for the clearance process? Should we be vetting in a different way? Should we have more people cleared? Fewer people cleared? And how should we be monitoring them after a clearance is granted?

GELLMAN: Who wants to start on that?

VINCI: I have one—

CARDILLO: I’m happy to start.

VINCI: Oh, no, go ahead.

CARDILLO: I mean, it’s flat broke now. So you can almost not do any worse than we’re doing today. It’s just broken. And, by the way, that’s not a new fact. Since I’ve joined it’s been broken. I think we need to get much more dynamic, meaning, you know, obviously the bank that issues my credit cards knows when I’m doing something out of the norm—(laughs)—because I get a text, or an alert, or call, and whatnot. We are so far from having that capability with our cleared population, it’s ridiculous. We tend to—and I know we’ve moved to a continuous, you know, cycle of evaluation, but those are just words. Every five to six years, right, somebody comes to visit me and asks me if I’ve done anything wrong in the past five or six years. It’s nuts.

So what am I trying to say? I think we should blow it up and start over—(laughs)—and maybe hire one of these, you know, financial tech firms to be our clearance process going forward. And I know that sounds a little flip, and you—and I know the government will never do exactly what I just said. But to your point—so I guess my answer to your question is, I think we should become more open, take more risk, and then put more agile monitors around it, such as we find in fintech.

GELLMAN: There’s an online question waiting.

OPERATOR: We’ll take our next question from Steve Kaplan. Mr. Kaplan, please accept the unmute now button.

Q: Hi. Steve Kaplan. Formerly of the intelligence community. Hello, Robert. Nice to see you.

A small comment and then a question. It’s more comment is, given the extraordinary number of people who have high-level clearances, it strikes me that it’s quite amazing that we have so few large leaks. The number of people with high clearances is just astounding.

Now, the question I have is, I’d really like to get a sense of where are the sharing problems today? Anthony talked about the third revolution after 9/11. I know it wasn’t Happy Valley after 9/11, because I was there. But I’m wondering what you think are the problems today? Where are the problems that you might want to flag?

WEINER: I think this whole conversation is reflective of perhaps some catch-up that needs to be made within the intel world, which is still very wedded to secrecy and classification. And it’s less about reluctance to share, because that’s different now, especially within federal agencies but increasingly from federal agencies to state and locals. There’s still some room for improvement in that domain. But it’s how much can we learn from the private sector, from all these other sources of intelligence that we haven’t necessarily tapped into, to the degree that we both really should and need to given the nature of the threat and its evolution. So the main problem is that we are not doing enough listening to the partners we need to listen to to help us keep our country as safe as we can.

VINCI: If I can say something on it. I think that if you—if you travel back in time to 1998 or 1999, and you’re in the FBI, and you asked the CIA for some information, they would say, go pound sand. And that led to a major problem. And we realized we needed these government agencies to share information, to work together. And I think what’s happening now is the threat is much wider now. It’s political, military, but it’s economic, it’s genetic, and involves health care, and it’s information operations targeted at regular people. And that means you need the intelligence community to work with people who aren’t even part of the government, who are companies, who are nonprofits, who are regular people.

And so this sharing problem now becomes even harder because it’s how do you decide what to share with essentially anyone? And how do you—how do you handle insider threat in that case? Or how do you handle that—it’s true in those organizations, there are almost definitely going to be foreign nationals, and maybe some of whom are, you know, under the thumb of an adversary in some way. And so we will have to figure out some means to handle classified and other intelligence in a spectrum that goes all the way out to individuals. And that, to me, is probably the hardest nut to crack, but the one that we have to work on because it’s the only way to tackle the threat.

GELLMAN: In the room, please. Sir.

Q: Hi. Jeremy Kutner from ProPublica.

I’m curious how capable you think U.S. industries are to actually counterintelligence collection on—sort of from foreign actors, particularly in light of, like, Salt Typhoon, and the like.

GELLMAN: In light of?

Q: Salt Typhoon.

GELLMAN: Right. I don’t know, Robert, how about I assign you that question?

CARDILLO: We’re amazingly good. I know I’ve come across as, you know, oh my god, you know, we have all these problems and these issues. But, to Anthony’s point earlier, I do believe America and our allies are very well-served by our intelligence community. I mean, I’ve lived the second, third, I hope to live the fourth revolution here, if Anthony’s book can kick us into gear.

GELLMAN: Sorry, Robert, can I just redirect that? The question focuses on how capable U.S. industry is countering—

CARDILLO: Oh, yeah, OK. Yeah, yeah, OK, I missed that part. But same comment. Our industry is amazing. Now, I’ve, you know, been more connected there for the past six years. And a day doesn’t go by that I’m not just blown away with the innovation and the—and the drive, and the risk calculus that people and companies and venture capitalists are willing to take. Which leads me to the—you know, I was still serving during kind of the real rise of China and the pivot and, you know, oh my goodness, they’re not only coming, they’re going to surpass us and whatnot, because, you know, all the projections of their GDP and their growth, you know, in 2010 and so. And, you know, my response was I have no interest in out-China-ing China. Anthony talks about—in the book—about how they, you know, use their political system to, you know, drive, you know, their nationwide goals.

I want to out-America China. And so I want to use that strength that the question just talked about, but do it in a way that that amplifies it. And, look, I know, you know, every administration comes in and says, we’re going to fix the acquisition process, right? We’re going to fix the clearance process, and whatnot. But it does take both. And I guess I would just ask us to consider even placing a greater bet on industry. You know, with the right oversight and with the right, you know, left and right bounds in our liberal democracy. But if we do that, I’m very, very bullish about our future.

GELLMAN: Thanks. There’s time for one quick last one sort of short question, short answer. Who’s got one? Well, in that case, I’m going to thank our panelists and wrap it up. A reminder, this was on the record. And we are going to close in time. Thanks so much, Rebecca, Anthony, Robert.

VINCI: Thank you. (Applause.)

(END)

This is an uncorrected transcript.