At the Science Diplomacy 2016 conference hosted by the American Association for the Advancement of Science (AAAS) Center for Science Diplomacy (publisher of Science & Diplomacy), Rose Gottemoeller discussed the role of science and scientist within disarmament, arms control, and nonproliferation.
Rose Gottemoeller: You know, AAAS is the very embodiment of science in diplomacy, and diplomacy in science; I was reminded of that [as I was] looking at the [presentation] slides flipping over. Science and Diplomacy magazine is actually a favorite reading material around our house, so I recommend it to you if you haven’t had a chance to look at it yet.
I was asked to give my view today on the role that science and scientists play in international relations, especially as it relates to my sphere of work—disarmament, arms control, nonproliferation. And I am delighted to have this opportunity, because there is no better place to make my appeal for help with my special problems.
So what are my special problems? As you’re aware if you’ve been working in the disarmament field for some time, the tradition began in the 1970s of monitoring and verifying that the arms control treaties and agreements were being properly implemented. Actually, I timed the launch of onsite inspection to the Antarctic Treaty, which was negotiated in the late 1950s. At the time, many people don’t know that the Antarctic Treaty was originally an arms control treaty because we were concerned that the Soviets were going to put ICBMs [Intercontinental Ballistic Missiles] in Antarctica and get us from behind, so to say. And the US had concerns about militarization also taking place. So the Antarctic Treaty became also, in addition to what it is today, a very powerful international legal platform to support science cooperation [and] environmental cooperation. It also began its life as a disarmament treaty with onsite inspection as part of the Antarctic Treaty’s implementation mechanism.
So fast-forward to the 1970s and our first negotiations with the Soviets at that time was, the SALT [Strategic Arms Limitation Talks] agreement, and we agreed [that] we would not interfere with each other’s national technical means; “national technical” means large satellites, high-flying aircraft, over-the-horizon radar capabilities. We would not interfere with national technical means that could be used to monitor large objects—ICBMs in their silos, bombers at their bases, submarines sitting at wharf-side in their ports. [We were really interested in] large objects; we were focused on the delivery vehicles, because we said if we can monitor the status of the delivery vehicles, we will be monitoring the status of the warheads. That’s all we needed to know at that point.
As time has gone on, though, our monitoring and verification problems became more difficult. We realized by the end of the 1970s that we needed onsite inspections in the nuclear disarmament realm, as well. It was first incorporated into the Intermediate-Range Nuclear Forces Treaty and now [it] is an old standby; it’s an important and very valuable aspect of the new START treaty [Strategic Arms Reduction Treaty]. But our problems are moving now from monitoring and verifying large objects—delivery vehicles—to monitoring and verifying small objects—warheads; warheads [that are] even held in storage facilities, [and] warheads not mated to their missile systems. Small objects, that’s one of my special problems. How do we monitor and verify small objects?
The second of my problems is how do we monitor processes? Processes. And a perfect example of that is what was agreed in the JCPOA [Joint Comprehensive Plan of Action]. The recent agreement with Iran had to do with monitoring their enrichment of uranium. Going forward [the treaty] ensured [that] it didn’t pass the threshold into high enrichment. That was a core aspect of the JCPOA. But monitoring enrichment processes is a challenge; it is a challenge to my community and one that, again, I look to your community to help us with.
I’m going to return to the Iran deal and that process, monitoring in a moment, but let me move on to talk a bit about how I’ve been looking at this problem over the last couple of years. The last time I stood on this podium, I had the opportunity to speak about my interests in the products of the information technology revolution and how they can help us with these two special problems of monitoring small objects and [the] monitoring process. I’ve been thinking about this and speaking about it.
Some of you may have heard me speak about the ways that we can use ubiquitous sensing to help us with these particular problems. Already they are in place, and many of you no doubt have one on your iPhone, radiation detection systems. Monitors can be downloaded as an app to iPhones [that are] already being used around the Fukushima power plant site in Japan by [the] local population to monitor radiation levels there. [The app] feeds into a centralized database and helps to enhance [not only] local knowledge [and] domestic knowledge, but also the national knowledge, and ultimately international scientific knowledge about what’s going on at Fukushima. So that’s one example of a sensor system on a dispersed mobile platform [that’s] already being used in practical ways.
Another example that I brought up when I was here last time is the fact that all iPads, all mobile platforms of that kind, have accelerometers in them; they, of course, are used to tell the device which side is up and which side needs to be facing the reader, but they also tell you if there’s [a] seismic event around. And you can imagine, as I said last time I was here, a whole network of such iPad accelerometers [are] somehow linked together and [they are] providing data for earthquake detection, but they could also be used potentially to monitor for illicit nuclear tests. Of course, you have to have additional analysis to distinguish between a naturally occurring event like an earthquake and an illicit nuclear test.
But those are some of the areas that we’ve been looking at, and [we’re] considering [them] as potentially interesting [tools] for helping us with our arms control monitoring problems. I have thought about it in two ways. One is help with the inspection problem, per se; how do we get more and better data to help us to identify small objects, to count and to characterize them, to monitor processes and to characterize them? So that’s one side of it, data for monitoring purposes. But the other side is very bread and butter: tools for inspectors.
I often like to tell the story of the early days of onsite inspection, and how in those days out in snowy Siberia our inspectors had to know how to ski. They had to know about cross-country skiing, because in order to ensure that there weren’t other [open] gates around a facility where missiles could be coming and going, they would ski the perimeter and just check to make sure that there were no other open gates. And then they’d proceed with their inspection. Well, you can imagine how time-consuming that was. They all had to be in really good shape, though, and I’ve heard many stories from these guys and gals, because some of them were women, and pretty short women, but they sure knew how to ski.
Anyway, long story short, you can imagine a tool for an inspector that makes use of some quite simple and now well-known software, such as Google Earth, to give a picture of a perimeter fence and [that] helps an inspector to understand in real time that there are no additional gaps in that perimeter fence. [There would be] no need to worry about missiles coming and going during an inspection process, [there would be] already a tool that helps the inspector to cut down on the amount of time he has to be at a site, and make his inspection more efficient. So that’s just one example. But it’s another area that I’ve been thinking about, again, in a very kind of bread-and-butter practical way. [It] would be very helpful to our world.
There have been, as I’ve carried this discussion forward, a big debate about societal verification and the problems of identification of individuals in arms control verification. And, of course, one can imagine that the Russians or the Chinese would not want their public involved in helping to track what’s happening with their mobile ICBMs, for example. No, this is a no-brainer. I take the criticism, I understand it, and we have to wrestle with both . . . well, we have to wrestle with the diplomatic aspects of this, the negotiability questions—would any of this be negotiable?—but that’s my problem, that’s not your problem. But we also have to then deal with the issues of privacy, issues of legality, and so forth, and I recognize and grant that. But I would just, at the moment, like to concentrate on and think about the applicability of these promising technologies, ubiquitous sensing technologies, particularly to the arms control and nonproliferation verification task.
[When] I think in terms of problems with identification of individuals, I can imagine a scenario where a country, wanting to convey to the entire international community that it is complying with a total ban, might pull its citizens in and say, “Can you help us to establish our bona fides in the international community?” For example, [the country would want to try to prove] that we are completely in compliance with the Chemical Weapons Convention, which is a total ban on the possession, production, deployment, etc., of chemical weapons. So [with] a total ban, I can imagine a government wanting its citizens to join in a societal verification effort to prove to the international community its compliance with that ban. But that’s the only example I can think of.
Otherwise, we do have to concern ourselves and consider anonymization of data, and I know this is for any number of reasons related to the privacy debate overall, but it is also a topic that’s being taken up by the scientific community and the IT [information technology] community. [It is a] very, very important set of questions and well worth [it], I think, to bring into this discourse about how such tools might be developed for arms control and nonproliferation purposes.
The last thing I’ll say about this as we consider it, these kinds of approaches have been a no-brainer for the environment community for a very long time. A crude example that I like to cite is the Gulf oil spill, already some years ago now, where local community activists got together with the local population down on the beaches of Texas—of Galveston and that area—and asked people to begin taking pictures and sending them into a central database to help the local environmentalists to scope the extent of the oil spill and the damage to Texas beaches.
This is an example that’s all over the place; the environmental community is doing this quite broadly. And at the strategic level in implementation and working out implementation of the Paris climate agreement, these kinds of approaches to monitoring carbon emissions are being considered, many of them big data approaches—I’ll come to that in a moment—but also involving internationally the public in monitoring and verification. So I just want to make that point.
On the environmental side it’s been pretty much a no-brainer. The Chinese didn’t like it a couple years ago when we put a pollution monitor in the rooftop of our embassy and started putting the data out on the embassy website every day, but eventually they decided they might as well go along to get along, I guess, but [they then] began joining in themselves [by] putting out more accurate environmental data about what’s going on in Beijing with the pollution. And, in fact, Xi Jinping and the Chinese government have been right at the center of leading the Paris climate accord efforts, and that’s been a very important step forward; it’s for the national interest and economic health of that country. Of course, countries only join in when it’s in their national interest to do so, but I think the case has now been made.
So leaving that topic, I wanted to expand the discussion today into another area that has been taking shape for some time, but I know is, again, not something that has been thought about in the realm of arms control nonproliferation verification and monitoring, and that is the Industrial Internet of Things. My eye was caught a couple of weeks ago by a very good article in the Financial Times, a report of Satya Nadella’s visit to the Hamburg Fair to talk about Microsoft’s interest in utilizing core data [that is] being thrown off by machinery to improve efficiency, safety, shareability, interactivity, performance, you name it. Of course, this is an area that’s taking shape; there’s a competition taking shape in the IT community about who’s going to be the winner. The article also mentioned Westinghouse is playing, and playing hard in this arena.
But I was attracted to it because of the way you could use data from industrial facilities, of course [this also means] adding in data from external sources, to optimize a task to optimize performance. I liked the example very, very much from the [Financial Times] article, [which spoke] of combining data from an individual aircraft’s performance, including its engine’s efficiency, with weather data along its expected flight plan in order to tweak the individual aircraft’s performance to enhance efficiency, but also [the possibility] to send it onto the route that might avoid some of the more difficult weather patterns to go through. So those are examples, I think, that are extremely interesting about combining external data with internal data to make a task more efficient.
So how could this apply to the arms control and nonproliferation space? I wanted to bring us back again to the Iran nuclear deal. The Iran nuclear deal contains a very important innovation; that is monitoring of the process of enrichment in Iran, again to make sure that the Iranians do not move again into the realm of producing high-enrichment uranium. It is at the heart of the efficacy of the JCPOA and [it is] extraordinarily, extraordinarily important to its long-term success over ten to fifteen years. This is an innovative approach, but it is not the first time it has been tried.
And I wanted to let you know that the roots of this effort are back in the monitoring regime that was put in place for the so-called HEU [highly enriched uranium] Purchase Agreement. We put in place at the enrichment facilities in Russia actual hardware on the enrichment lines to monitor the level of enrichment. This was a transparency measure, different from monitoring and verifying an agreement, because when you’re verifying an agreement there’s an obligation in the agreement and you’re monitoring to ensure that they are reaching and implementing their obligation.
A transparency measure, as it was, in the HEU Purchase Agreement was just to make sure that the quality of the enriched uranium was at a certain level, and so in that way from time to time, we would have our experts go in and read the data. But it was the fact that early on, and beginning in the 1990s, we were beginning to monitor enrichment with the cooperation [by] working together with the Russians at several enrichment facilities in Russia. So this threshold has been stepped over, now how do we make it more efficient and effective? And [I think that] the Industrial Internet of Things, is [the] way [to] very seriously go about it.
I can envision—and I’m going to wrap up here and move to our discussion—but I can envision two ways particularly that the Industrial Internet of Things could help with the problems that I have, first of all, so-called safeguards by design. In the Safeguards Agreements we have long had ways to work with countries to ensure [that] they have access to facilities [that] inspectors can go in. The notion of safeguards by design is that when a new facility is being built, you build it to optimize the efficiency of safeguards activities to ensure that inspectors have easy access to what they need to say. To ensure that there are certain monitoring capabilities—CCTV [closed circuit television] cameras is a simple example—[safeguards] are built in to ensure that there are no hidden nooks and crannies that would arouse questions, [particularly, in regards to] over [the] many, many years of the performance of a facility in a safeguards regime. So safeguards by design, I can imagine, would be a very relevant area to look at the applicability of big data [within] the Industrial Internet of Things, [in regards] to the nonproliferation safeguards problems.
The other is related—and, again, I welcome from this audience perhaps some additional ideas—but the other is also related to the civil nuclear power issue going forward, and the insurance that a civil nuclear power program does not bleed over into a weapons program. And, of course, they are of necessity Siamese twins, they are joined together. But we have to look for ways, and the community has always been looking for ways to introduce a separation. One way is proliferation resistance in reactor design, and here again the applicability of the Industrial Internet of Things and big data could be a way to enhance our ability to design nuclear reactors that are resistant to proliferation while optimized for the production of energy, and perhaps optimized also for the minimization of waste. [There are] many, many opportunities and ideas out there.
So thank you very much for your attention this morning. I look forward very, very much to our discussion. I am quite serious about working with the scientific community on these important problems and have spoken recently in Silicon Valley. We had a very interesting workshop back in the first week of April, drawing together Silicon Valley companies as well as academic players. It was held at Stanford University [and there were] very, very interesting discussions focused on these set of issues. If you are interested, because the group that formed as a result of that workshop is continuing to look for practical ways to bring these kinds of ideas into implementation, and [this would include the] monitoring of arms control treaties and agreements.
So if you are interested in joining this, please be in touch, and we will definitely look for ways to link you up with the community out at Stanford and in Silicon Valley, because there are a lot of people thinking about this [and there are] a lot of people thinking about it in our national laboratories, as well. I see a couple of laboratory colleagues around the audience, and there’s some very good work going on already. But I welcome widening of the circle, and I’ve been very keen to have both the larger academic community here in the United States, but also the larger international community involved in this effort. So I welcome your comments, I welcome your questions. Thank you very much for your attention.