Q: What are your current research interests and ongoing projects?
Dahlgren: “In the missile world, I’m currently interested in missile defense countermeasures. This has largely been my main project outside of work. Broadly speaking, countermeasures are devices used spoof missile defense systems. While much of the current discourse focuses on how effectively countermeasures work, I am investigating historical case studies of how states acquire them and how they fit into states’ strategic thinking.
LaFoy: “I’m currently revisiting a satellite coverage simulation that I was working on in 2016 and 2017, when North Korea was blasting off missiles every two weeks. My basic question was, “If an analyst with a shoestring budget and enough commercial observation satellite imagery could simulate when the holes in the observation were, could the analyst then ballpark when there were no satellites looking at you? (Outside of stealth satellites).” The answer is yes! I’ve created an experimental output that tweets every time Sinpo port in North Korea is under an observation satellite. Basically, I want to be able to tell when things can be seen and when they cannot be seen in a simple binary (i.e. visible, or not visible). However, there are a lot of variables to account for… (such as cameras, satellites, and sensors), it’s ultimately difficult to actually tell who is listening or watching, for how long, and if they are even looking at the imagery or not.”
Q: What technical issues in national security interest you the most?
Dahlgren: “There’s a lot policy questions to unpack regarding the recent revolution in machine learning. I’m particularly interested in how these technologies could impact the discrimination problem in missile defense. Could new classifier technologies help missile defense sensors sort out real warheads from decoys and debris?
LaFoy: “Machine learning and AI has also been on my mind a lot lately.”
Dahlgren: “Yeah… it’s super popular in DC!”
LaFoy: “I have a particular fascination with AI and machine learning and ethics. For those that utilize and exploit Ai or machine learning, it becomes black-boxed because so much of the technologies are constantly changing. As these systems develop, it becomes very complicated to track what it does and how it responds to inputs and outputs. First, we need more education on the technologies themselves and the implicit biases that build datasets and potentially unethical algorithms, before they can be used more widely. I’m also interested in the OSINT applications, and how automatically collated systems, in a way, make operational and informational security very complicated.”
Dahlgren: “It’s very unintuitive how neural network-based programs reach the conclusions they reach; you can create a judge that can accurately discern the facts of the case but can’t tell you why it made the decision it did.”
LaFoy: “It’s hard to accurately evaluate the data unless you train and test it on a human…But do you want something more accurate than a human? I also think this issue leads into situations where data is being harvested from locations people think are traditionally secure (i.e. texts or documents hosted in Google Drive). It’s concerning how these data sets can be turned against the end-user. Additionally, specifically regarding AI and missiles, from a human intelligence (HUMINT) sense, people are constantly innovating with large data sets around them. I think AI and machine learning create interesting opportunities to automate surveillance.”
LaFoy: “Overall, I think AI and machine learning aid in processing large amounts of data and figuring out what is the noise.”
Q: How did you get into the technical area of national security?
Dahlgren: “I was very interested in nuclear issues in school, and anyone that has an interest in nuclear issues tends to be interested in how technology impacts national security. That, and listening to Scott and Jeffrey Lewis’ podcast, “Arms Control Wonk” — which got me fascinated with the field.
LaFoy: “So originally, I was more of a North Korea person, and interested more in the political theory and traditional International Relation side of things. I was really interested in the impact of nuclear weapons and North Korean policy as it relates to the international community. Eventually I studied nuclear weapons and ballistic missiles more in-depth to better understand the technical side. I am drawn to these weapons as a focus of study because they are unique tools of power and diplomacy. Nuclear missiles are a weapon of policy more than a weapon of war. The technical details and margins matter for treaties and military postures in a way that is rarely represented in other weapon systems. It would be rare for a new type of tank to totally obviate some form of ground warfare and require an adversary to totally change their military or diplomatic posture, but missiles have a much more direct effect on policy and reactions. They are the materiel directly affecting policy, the international order, how people get along, in a way that conventional systems don’t. They’re not an instrument of policy power, but they have huge policy implications. There’s also this unique US/North Korea dichotomy with regards to missiles that is fascinating.”
Q: How have you found ways to convey your technical subject areas to a less technical, potentially policy-oriented crowd?
Dahlgren: “I use social media outreach — especially Twitter — to talk about my research. This usually takes the form of posting a striking image related to my ongoing research and explaining the history and context behind it. Images grab people’s attention, and Twitter’s limits force me to write in a way that is digestible; in a way that might easily translate to social media. Using social media allowed my research to reach a broader audience than I thought possible, but more importantly, taught me habits of writing and thinking to make the rest of my work more accessible.