Kristen Wood spent 20 years at the CIA working on operations and analysis. She finished her career there helping lead the innovation and technology group at CIA's Open Source Center. Since leaving the agency, Kristen founded and led a company that provided open source solutions to the government. She is now serving as a fellow at the Harvard Kennedy School's Belfer Center for Science and International Affairs. Kristen has been on our show before, and she joins us today to talk about the potential for an open source revolution in intelligence. We'll be right back with that discussion after a break. I'm Michael Morrell, and this is Intelligence Matters. Kristen, welcome back to Intelligence Matters. It's great to have you on the show again. Michael, it is great to be here. Thanks for inviting me. You're welcome. So we are going to talk about something called open source intelligence. This is an extraordinarily important topic, and I know you agree, and thank you for agreeing to come on and talk about it. I'm really looking forward to the conversation. This doesn't get the attention that it deserves, and maybe we can help bring a little bit more attention to it. the place I want to start, Kristen, is to ask you to define it, right? When people say open source, open source intelligence, what do they mean? So your skill at asking the right question continues. And it's a problem. And starting at a fundamental place, like what is it? Because there isn't a universally agreed upon definition. There are pockets of the Defense Department that talk a lot about PAI and CIA, so publicly available information and commercially available information. And they say that when you do analysis on that data, then that it becomes intelligence. I think at CIA or at the DNI's open source center, open source enterprise, it's been the exploitation analysis of all sorts of data, classically media, TV, newspapers, academic journals, publications. And then really with the advent of social media and online media, maybe 20 years ago, it also incorporated those things. Fast forward to now, and the definition still uncertain and needs to broaden to include commercial information that's way beyond the scope of what open source classically used to be. So retail sales information, financial data, information on car sales, information that people share on their cell phones, online shopping information. And it goes on and on and on as the digital interconnectivity of devices expands across the globe. What's the difference between publicly available information and commercially available? So publicly available information is generally seen as that you could not Google because there are better tools available to those who are experts than what we would normally do to find a great Indian restaurant. But it is that that I can just find. I don't need to pay for it, and it is easily available to most people. Commercially available information is something that you'd have to pay for, generally. So maybe that's a company that has information about financial transactions, or a railroad company that sells its data on on-time arrivals of trains, for example. So those are generally the distinctions. although there's not complete agreement about that within the open source community. And then would you put the private sector doing analysis using both of those types of information? Would you consider that to be open source? I would. And I think one of, as we get into the conversation, one of the hallmarks of this moment in open source is that it is no longer just intelligence officers or military officials who have access to phenomenally important information for national security. And it's many of these private organizations that have remarkable data and tools to tell stories about national security that really we could have only dreamed about having a source to collect. So they're a big part of how this is all evolved as companies like Google and Amazon learn a ton about us. And as we share, frankly, a lot of information about our personal lives on platforms like Twitter and Facebook and for some TikTok and elsewhere. So it is a rich, rich set of data that is available to tell important stories. So I think we do have to count the private sector, academia as well, and the nonprofit world in this big community-related open source. Okay, Kristen, what's your background with regard to open source? So I think for my 20 years at CIA as an analyst, brief time in operations, time we worked together as a PDB briefer, I used it throughout my career. But for me, for the most part, it was a little bit like garnish on a plate. It was something to add color to the really hard-hitting intelligence that was either collected by a human asset, via signals intelligence, or perhaps via imagery. It's great at identifying crises, right? There's an earthquake, a hurricane, an act of terror, an act of civil disobedience, or maybe a leader announces that they're stepping down or they've taken over power. But in those cases, it's really just a starting point for professionals, intelligence professionals to pick up and do the job. My last job at CIA was as the deputy chief of innovation and technology at the Open Source Center. And I think in 2015, it had already changed considerably in terms of value. What was that job? What did you do? So it was really looking at how do we need to evolve from a classic infrastructure that was built for media reports and analysis, reports on news channels, publications, to how do you accommodate the Internet and the vast amount of data that's now available and coming in at speed. and in 2015 we were really in the beginning steps of that and I know that work has continued much since then as the challenge has as well and then when did you when did you retire just to put this in context 2015 okay what was the most impactful piece of open source information that you remember from your career well I'd love to ask you the same thing this is my podcast So I think for me, the most impactful was something I shared with Vice President Cheney and his national security advisor, Scooter Libby, right after the 9-11 attacks. And it was a piece that pulled together global reactions to the attacks and articulated the acts of solidarity that were happening across the globe. people lighting candles, creating a U.S. flag on the beach in Australia, lighting their national monuments red, white, and blue. So sentiment analysis that said really the world was with us. And I think at that moment, as you recall, too, it was such rage and grief. And this sense that we had to figure this out, that knowing where the rest of the world was coming from, I think was a boost to them and to the other people who saw it. So I also remember in the immediate aftermath of 9-11 that there was polling done across Muslim countries that showed a significant amount of support for bin Laden and for what he had done. So there was another side to that, right? To that open source information out there. You asked me, kind of jokingly, what the most important piece of open source was to me. And just kind of sitting here thinking about it, before the Bin Laden raid, we wanted to know as much as we could about the city of Abbottabad. And we didn't want to trigger that we were interested in the city of Abbottabad. so we asked the open source enterprise by that time right the name had changed since you left the open source enterprise to do reports on us for a pretty significant number of pakistani cities and abadabad was one of them and so we got back this package of reports that really provided really detailed information on Abbottabad that was extraordinarily helpful, not only to CIA, but also, you know, more importantly, to the SEALs who had to go in to Abbottabad. So, you know, that's kind of my answer to your question. You know, maybe to some extent, Kristen, you've already done this, but from a kind of a high level, can you kind of walk through the history of open source in the intelligence community, and then where are the big inflection points there, right? I think you've touched on this a little bit, but maybe just let's just go through it one more time. Absolutely. I mean, it started as the Foreign Broadcast Monitoring Service in 1941. Oh my gosh. At the time, it was under the Federal Communications Commission. And it was stood up to monitor access powers and their propaganda through shortwave radio. You know after Pearl Harbor they changed the name still under FCC to the Foreign Broadcast Intelligence Service And they started monitoring official stations in many countries and black stations interestingly that weren what they pretended to be And these were broadcasting attacks on President Roosevelt at the time, while pretending to be stations that were in the American Midwest and focused on the Midwest issues. And the tactic they were using was to stir up racial tensions and other issues. And as I was thinking about the other day, I thought, gosh, that sure sounds familiar, doesn't it? Yeah, 2016. So in 1946, it was renamed FBIS, the Foreign Broadcast Information Service, and became part of the new CIA as part of the National Security Act of 1947. And fast forward a lot, in 2005, it became the DNI's open source center. And that was in response to the WMD Commission's recommendations to devote more attention and resources to exploiting openly available information. And as you know, the director of CIA then became the functional manager. And their mission expanded to include Internet, databases, video, geospatial data, lots of other stuff. And in 2015, it became OSE, as you mentioned, Open Source Enterprise, and was moved to the Directorate of Innovation at CIA. And the big inflection points are the internet and then social media. Is that a fair way to think about it? Or are there more data points there? I think those are the big ones. radio, shortwave radio in the beginning, the development of the internet, online media, social media. And then I think this place where we are now, where the digital interconductivity devices is really almost the technical definition of ubiquitous. Most of us count our smartphone as a limb at this point. And we do a lot of our work on them. And so the part that I think OSE is grappling with now is how and what of that enormous body of data needs to be pulled in for the national security mission. So back to this arc of this history for a second, how has, in your view the intelligence community kept up with the opportunities that are available in the open source world? And I don't want to prejudge the answer, but I'm going to kind of do that a little bit here. Has the gap between those opportunities being there and taking advantage of them, has that gap grown over time? So this is a tough one. And I guess I want to start by saying I'm really hopeful that more transformation is coming. Randy Nixon, who we both know well, was recently appointed as the director for OSE. We both know he's a terrific officer and he understands the new world we're in. But I think there was something really telling, Red, that former NSA and CIA director General Hayden recalled as a turning point for the United States for American intel around the mid-90s. And that was really, there was a debate about whether we would focus on cyber dominance or dominating the information sphere more generally to include diplomacy, public affairs, disinformation, things like that. We decided, he says, with much debate to focus on cyber dominance. And the Russians went to information dominance. And we're seeing the impact of that everywhere. So as much as it pains me to say this, We just haven't seen major shifts to accommodate the world that's existed for the last three or four years in particular. There have been many, many IC studies, intelligence community studies about what to do about open source. I'm sure you've read many, many of them. And a lot of recognition on the Hill within leadership and the IC and elsewhere about how important this is. But the necessary urgency just hasn't been there. We're seeing the effects of the U.S. not having the window it needs into what is happening in the open world as it's affecting the underpinnings of our democracy. We mentioned the attacks, the U.S. presidential elections, but also the attacks for misinformation and disinformation that are eroding faith in our institutions. And so I think we have much to learn here. We did see that I think there's been a frustration with OSE across DOD and the IC because the mission is so big. And it sounds really critical to say, but I think about it, NGA has published National Geospatial Agency, which does imagery. They have about 15,000 people on the mission. NSA has more. If I think about OSE having that same functional responsibility, well, I don't want to get into numbers. It's just tiny. And so, and both in terms of staff and budget. And so in the world we're in, there has to be a sea change. And I think as a coping mechanism, if you will, many, many, many entities within other government agencies, within DOD and elsewhere have stood up their own open source efforts because they see the urgency of it and they are trying to get to the heart of what they need to learn for their missions. And the shift that we need really is to bring this all together to create more of a, like there's a human service and there's geo-int and signals and there really needs to be that kind of move forward for open source now. We're going to take a quick break and we'll be right back with more of a discussion with Kristen Wood. So Kristen, what I would love to do for a few minutes here is get some examples of what it is we're talking about, right? So maybe if we could take an example of publicly available information that might surprise people in terms of that information being useful to intelligence officers, and then maybe commercially, an example of commercially available information, and then an example of somebody in the private sector who actually takes that information and turns it into an assessment, right? Turns it into analysis, turns it into something that's useful to decision makers. Can we do that? So I think to start with the publicly available information, it's really interesting that as COVID just started to take hold in our country, that some of the greatest guidelines or guide posts for both symptoms and where it was spreading came through Google searches. So people typing in, you know, here's my list of symptoms, people asking what diseases or what illnesses are associated with a loss of sense of smell or taste or bright red toes. And for those who map that kind of information, it was really invaluable to see where the COVID was progressing and how prevalent it was in many places. And in some cases, actually in many cases, the organizations that pulled up that information were well ahead of the CDC in understanding both symptoms and spread. I was actually doing that myself, right? I was going to websites and collecting data and keeping track of it because I was just interested, right, in what was happening. So I was actually using the information that you were talking about. I'm not surprised. I'm not surprised. I'm weird, I know. No, no, no, you're dedicated. Um, yeah. So commercially available information, so information you can buy to create insights. And much of this has to do with what we call person in quotes information. So that can be tracking your iPhone. That can be your car's telematics system, sharing where you are, how fast you're going. It's a problem for me. And other data about where you stop. And so that information can be pulled together and has been pulled together to understand more about terrorist groups. We saw it throughout the January 6th hearings and the released data. There's a ton of that available to show how those, that protest, those attacks were planned out in the open. And so a lot of the data I think that the Bureau has used to find those people has been based on that kind of data, that they obviously have the right authorities and permissions from the courts to use. So I would say that would be a good example of commercial data. so in terms of can you then take that data and anonymize it right so that you're not focusing on specific individuals but you but then you start looking at big trends right across that data and there's some really remarkable insights you can get where actually knowing who the person is or what their specific characteristics are gets in the way on those big trends it's really looking at things from where are people moving and so something as simple as migration patterns right so with climate change particularly in latin america you see an enormous movement of population in areas that just aren't either farmable or habitable so many many ways that can be used but the one that is the one that comes to mind right now okay what about analysis so Well analysis is really fun in this sector because you dealing with data that is shareable So we'll talk about the insights and then maybe what it offers us next, but the insights are remarkable. There's a financial company that does, they have looked at the DNA of every company there is that's publicly traded anywhere in the world. And they have pulled together through their AI financial AI modeling, pulled into it every bit of data about those companies, whether it was video, audio, text, and it has allowed them to create the DNA of companies. And their whole reason for being is to work with investment firms and very, very serious investors to help them improve their outcomes. And according to their CEO, they perform on a par with Warren Buffett and all AI and data driven, remarkable work. And in conversations with them, I said, well, what can you tell us about issues of national security concern? And two things, we went to Russia and they're able to look at the Russian economy across all of its publicly traded companies and address the real impact of sanctions. And it's both wildly granular, company by company, its location in Moscow versus somewhere else, and also deeply strategic. So they can tell you how much the Russian economy contracted as a result of sanctions, how much it's a bargain, considered a bargain now, but you're having where you would normally have bargain hunters come in. You're seeing people sit on the sidelines because they're not really sure what's going to happen within the war. And then obviously with the leader of Russia. So that's one example. Another is, and this is both important for us offensively and defensively, is they can see when companies aren't normal, quote unquote, when their revenue doesn't match their sales, when their board members are board members in multiple companies, or they win U.S. tech funds, and then they fold. And something with a similar name competes for U.S. tech funds elsewhere. And you'd be shocked to know that a whole lot of those companies originate in China. so they don't that has nothing to do with their reason for being but they care very deeply about democracy and preserving it and they were able to step in and provide that information data driven very very hard to challenge and i think it's both useful for us as we're looking to make sure those companies don't get our tech funds and maybe the fbi would like to know that information as well, but also as maybe we're doing things overseas that we wouldn't want people to know about. You know, the entities that do this analysis, do they organize themselves by a particular substantive focus, a particular source of information? How does that work? So I've had the opportunity to talk to about 200 companies in the last few years that are in this space? And I would say the answer is both. Many of them are analytic tools companies. So they have artificial intelligence, AI, ML, machine learning, or natural language processing, NLP capability. And they have either purchased data or brought it in from the internet or developed them themselves. And so some of them will are really, they have the data to create sales for their platform, while others have the data, then they're really not interested in creating a platform around it. So, I mean, you have a company like Flashpoint that does remarkable work in the space in terms of helping people, organizations understand what's happening in the world that's both platform and data. We're going to take another quick break. We'll be right back with more Intelligence Matters. Stay with us. Here's maybe an unfair question or a tough question. Maybe it's unfair because it's tough. If you think back to the information that you had as an analyst on the various different issues that you worked during your career, and you think about the role that open source played at the time and what kind of percentage of the unique information that it brought you at that time, what would that percentage have been and what could be, right, if it was fully used, what could that percentage be today? In other words, how much of the intelligence mission could be supported, I guess I'm asking, could be supported by open source intelligence? it was probably a much more articulate way of asking that question and you don't seem to ask the easy ones so consistent with your skills so i guess the answer for historically is i i do not remember a single report and i'm not saying they they weren't there i just don't remember anything that's notable in probably the first 15, 16 years of my career. And it's not that there wasn't interesting data, but it wasn't central to my mission. You know, I think about when we talked recently about the Iraq war and the terrorism question is, had Saddam had a role and 9-11, and we didn't have data. So if we had been able to get phone data from Abu Musab al-Zarqawi, you know, a time leader of Ansar al-Islam, and his cohorts, not only in northern Iraq, where they were positioned, but when they were in Baghdad or elsewhere, and we were able to map out their network with the links charts that people have been able to do since then, I just think we would have had much a better ability to provide better answers to the policymakers. And so if at the world now, I think there is a tremendous change possible. I mean, the United States Intelligence Community, the government provided collected information. So signals intelligence, human intelligence, national technical means, mass imagery, has really been the bread and butter of the intelligence community since 1947 and before, right? But the world has changed, and open source is never going to replace that. But it could become the int of first resort. Why would we want that? Well, we might want it because it's less expensive. If we think about launching satellites and processing capabilities, that are required to bring something down for an analyst to review. That can be expensive. The same thing about signals and others. A human asset, we have to train a case officer and have them trained in language and then send them overseas to find someone to collect the information. Wildly expensive, both in terms of risk to people, but time. And so it has the capability now to become our INT, our intelligence tool of first resort to inform classified systems and information, or frankly, maybe it will prevent us from having to do something we might have had to do otherwise. But it sounds like you're saying, and tell me if I'm wrong, but it sounds like you're saying that maybe when you and I were new analysts, open source played a small role, and you're saying that if used to its fullest today, it could play a much, much, much more significant role. Oh, absolutely. There's a lot of much. It is, but they're all valid. No, I think that that is the truth. And the data is available instantly or nearly real time. And it's shareable. So I think about, you know, maybe we have a really sensitive asset who's provided us information. And we can collect, we can get open source, buy some open source data that tells the same story, right? We're not exposing that person and, and we're risking them with others. So there's, there is a, this is a sea change moment. And I'm deeply, deeply concerned about how far behind we are, our adversaries and taking advantage of it. I want to do two more things, Kristen, in the time we have left. The first is, I assume you would agree that there remains a role for the traditional ints, right? There are issues that you actually need a human source, or you actually need NSA to get the right penetration to learn some information that is still not available through the mining of open source. Is that fair or not? And I don't want to create the impression that I think otherwise. It's vital. But perhaps we can take the pressure off by allowing people to spend longer focusing on those exquisite operations. And perhaps there's something that we don't have to do because it's available publicly. I think it's much easier to find information now in open source, and it's also harder for us to hide. our activities. And some of our most sensitive emissions can be inadvertently exposed. And I know that's not the topic for today. But I think until we understand the space as well as we need to, we're at more risk, I think, than we would want to be. Okay, so here's the last theme I want to explore So if we not where we need to be today in exploiting open source why aren we and what do we need to do to fix it And I guess part of that question part of the second half of the question is some people have argued that we need to pull the open source enterprise out of CIA. How can an open source enterprise thrive in an organization whose fundamental mission is to collect secrets and create a separate agency within the IC or even outside the IC. So what's preventing us from taking full advantage and what would you do to fix that? And would that include the creation of a new organization? You do ask really simple questions. I think there are, to me, five key challenges. One is the culture of secrecy. And it's been there for good reasons, right? We all have security clearances. We live in secure compartmented information facilities. We read classified information all day on classified systems. And most of the people we talk with have clearances. And for our entire history, it has been the richest source of data available to us to provide analysis for policymakers to create decision advantage. And so I think that has created a completely understandable bias towards and a familiarity with secret information. On the flip side, because we live in these closed worlds, we don't have the same intimate film familiarity with the open world. it's hard to know the space if we are only visiting it through specific computer access and occasionally for a few hours a day or even all day but we're not able to peruse it the way that it lives so i i think we're not too far from a point where there are going to be very few secrets and the name of the game really is going to be speed to insight for decision advantage. So it's that pivot that we aren't there yet. Acquisition and the slow pace of bringing in tools and data. And this is something I'm sure many, many, many of your podcast visitors have talked about, but it also affects open source because it can take a year or more to get something on contract and then getting it into the classified system can be tough. So it doesn't support speed to mission. And then I think the scale of the challenge versus the organization, and we talked about this briefly, NGA with 15,000 employees, NSA with more, OSE is microscopic by comparison. And the data they have, the challenge, the zettabytes of data that's collected is equal, maybe even expanding faster than that that NSA and NGA need to address. So expertise and I think this one hurts. The IC doesn't own this space with some notable exceptions and they really are remarkable. Individuals and small units the real expertise is out in the open world. I mean companies like McDonald's and Starbucks and Disney and Dow all have open source units to help them understand the world and protect their brands, and they don't have access to classified information. But they're able to make billion-dollar decisions based on information on the open. So expertise on this new, bigger commercial world is outside of government. And companies like Flashpoint, as I mentioned, Institute for the Study of War, actually, it's a nonprofit that has the best maps of the Ukraine conflict, and it's all based on open source. For sure. So do we need a new organization? Yeah. So very smart people have said we need a new open source agency. We definitely need something very serious, very fast. And here's my concern. An open source agency can't happen fast. Congress would have to pass legislation and fund it. New leadership would have to create priorities and strategic plans and mission statements and find space. And the ISC agencies and probably DOD would be taxed to provide part of their budgets and send many hundreds of people to the new agency. And then they would need to get moved and trained and provided tools and data. And that's all going to take years. And I don't think the world is going to wait for us to supersize what we do and empower it with enough staff and tools and funding to compete. So I think for me, the solution is a little, maybe that begins, but the solution is a little bit to me more like an archery target. So if you imagine those concentric rings of opportunity and open source right now, the current IC is in the bullseye. The units do open source across the national security space. They're usually embedded in the mission and cleared and they need to be. You don't want the same China expert on an aircraft carrier that you would want at the FBI's intel shop or maybe supporting CIA targeting. All those entities have different authorities and missions and you need open source and mission expertise. But we don't have the standard tradecraft training and career field across all of those. And so it's an area I think we need to get really humble about asking for help. So, you know, beyond the bullseye, I think that first ring is about how do we get to speed as if an open source agency is decided upon, how do we get to speed and what and how do we outsource. And so some have suggested it could be like a Manhattan project. I can imagine, I mean, you remember when In-Q-Tel was stood up, government funding, it operates independently, but the government funds it every year. And they were given a set of priorities and told to go make it happen to provide tech capabilities, to fund companies that could provide tech capabilities to government missions. And here in this kind of an entity, you wouldn't need a clearance or to operate on the high side. And I think you could keep living in that unclassified world would allow them to move them as quickly as internet does and build a consortium of companies. Don't pick a winner. Don't pick one AI company or one NLP company or one data provider, but create a consortium and partner with industry the same way the geo-int community does with commercial imagery providers. It would be very easy to vet them from malign influence and set them to work. And I know there's so much willingness across serious companies, even competitors, to work together because they're so concerned about the future of the country. And then I think the second ring is really insights from companies and entities that aren't traditionally about intelligence. I spoke about one investment company, but venture, private equity, transportation, manufacturing, retail sales, just really don't know how useful their information could be to government. But I'm betting a company that's about to make a multi-billion dollar investment in another area of the world knows a lot about its government, who's in power, prospects for unrest, et cetera. And then I think the next ring would be think tanks and nonprofits. You know a lot about those, but how can we get all their published works a little bit like OSE has done historically, but use AI to highlight what is needed at the speed it's published. And then I think the last one is one we haven't really touched on yet, which is kind of the toughest to get our arms around. And it's the field of open source investigators. Bellingcat is the most well-known entity that does this open source investigation, but these are not people necessarily with any expertise. They just have, they're deeply passionate about getting to the bottom of things. And they may not even be college graduates or Americans, but there is really remarkable, highly rigorous, transparent tradecraft being developed in this space that's some of the best in the world. And they have uncovered really remarkable things we would call secrets, like who was behind the downing of the Malaysian airlines over Ukraine, who was involved in the poisoning of Skripal in England, the Russian dissident. right they discovered that gru illegals were involved in that and there are just so many more examples of these truly hard sensitive missions we would call if i'm sitting in a skiff i might even be in you know code word situation where you know it's an apartment and they're able to get that information now. So it really, it really sounds, it really sounds Kristen, like if you took your first suggestion, right, which was have the DNI say to a group of people, go figure this out and come back and tell me what to do. Just like we did with InqTel, right? That you could cover all of these pieces. So that's what I think we need to do. And I know just the person who should run it, it sounds to me. So I'm offering your name to tackle this problem. So I just want to say thank you for taking the time to spend with us. This has been a fascinating conversation. We've actually gone way over, but thank you. And this is an extraordinarily important topic and we should come back at some point and talk more about it. But thank you for taking the time to join us. Thank you so much, Michael, for your time. I really appreciate the opportunity to talk about open source with you. That was Kristen Wood. I'm Michael Morrell. Please join us next week for another episode of Intelligence Matters. Intelligence Matters is produced by Olivia Gassis, Jamie Benson, Paulina Smolinski, and Reggie Bazile. For more from this week's show, visit cbsnews.com. Intelligence Matters is a production of CBS News. Three judges, one bench, zero room for nonsense. I'm Judge Dan Mentzer. Joining me are Judge Rachel Juarez and Judge Odete Tewelde. And on Hot Bench, we don't just hear cases, we debate. What she's asking for is for the payments that she made on his car. But there's still payments to be made. Correct. And we deliver justice. That is the verdict of the court. Follow and listen to the Hot Bench podcast on the free Odyssey app or wherever you get your podcasts.
