The implications for GEOINT in the monitoring of international nonproliferation agreements
By Pia Ulrich and Chris Bidwell, Federation of American Scientists; John Lauder, Nuclear Verification Capabilities Independent Task Force; Harvey Rishikof, American Bar Association Standing Committee on Law and National Security; and Valerie Lincy, Wisconsin Project on Nuclear Arms Control
A vital task for the geospatial intelligence (GEOINT) community remains collecting, analyzing, and exploiting data for the monitoring of international agreements and informing verification decisions concerning compliance with those agreements. The fundamental challenge of monitoring and verification was a staple of Cold War intelligence—and one of its greatest achievements. This essential role has gained renewed urgency due to the demands of recent negotiations and compliance issues concerning Iran, North Korea, Syria, Russia, and potentially China.
During the Cold War, the United States and the Soviet Union sought to view each other’s nuclear delivery arsenals, conventional military forces, and industrial infrastructure from space. The U.S., in particular, built eye-watering intelligence capabilities for monitoring that made arms control agreements possible and reduced the risk of strategic surprise and miscalculation. International treaties and other agreements referred to these capabilities as national technical means (NTM) and prohibited interference in their use for monitoring. The agreements did not define specifically what NTM included. This intentional ambiguity provided useful flexibility among the parties as to what methods of technical monitoring would permissibly be applied. NTM was understood to include more than remote sensing from space, but satellite imagery was clearly viewed as a major component of NTM. These NTM were later adapted for use against a host of other national security issues and have continued to improve in terms of sensor types, spatial and temporal resolution, and accuracy.
Since the Cold War, a major evolution in remote sensing data has been the increase in publicly available data and the number of available observation platforms, including those operated by private entities. Less expensive technologies and new business paradigms yielded a robust industry and marketplace. The National Geospatial-Intelligence Agency’s (NGA) Commercial GEOINT Strategy observes that the remote sensing industry continues to evolve in terms of global coverage, rapid revisit rates, diverse spectral content, aggregation from open-source venues, and analytic capabilities with increasing dynamism. The impact on the nonproliferation and monitoring communities is threefold:
- The accelerating quality, quantity, and timeliness of imagery and other forms of remote sensing available outside of governments.
- The growing volume and availability of worldwide transactional data related to commerce.
- The ease of both accessing the data (including imagery) and communicating findings, observations, and assertions about illicit activities related to nuclear programs and proliferation (with varying degrees of accuracy and truthfulness) through an increasing number of traditional and emerging social media outlets.
Overlaying these developments is the introduction of new forms of data analytics, including artificial intelligence (AI) approaches such as machine learning, which serve to speed up both the process and pace at which these developments can affect monitoring and verification activities. The sheer volume of available data, imagery, and analysis, some of it conflicting, has made the monitoring process more challenging. In addition, confidence in the result of data analytics is moderated by a lack of understanding regarding the logic basis (algorithms) that produced the result and the extent to which it can be generalized. Additionally, there is a risk to data integrity due to a dependence on the digital cloud, data storage, web browsing, and online communication.
- This article is part of USGIF’s 2019 State & Future of GEOINT Report. Download the PDF to view the report in its entirety.
Commercial Imagery and NGOs
Enabled by these increases in the speed and quantity of open data sources, non-governmental organizations (NGOs) are playing an enhanced role in commenting on nonproliferation agreements, facilitating greater transparency, and helping to identify options, opportunities, and challenges. Use of enhanced open-source tools by the NGO community is likely to increase as the technologies continue to improve and costs decline. Dr. Christopher Stubbs of Harvard University and Dr. Sidney Drell of Stanford University, in “Public Domain Treaty Compliance Verification in the Digital Age,” described these new tools collectively as “public technical means (PTM).”
One significant change that has emerged from less expensive and more accessible geospatial information is the emerging private sector business market. Unlike legacy aerospace firms that focused on national security clients, many new private firms are financially incentivized to sell their products to as many customers as they can, including businesses, foreign governments, and NGOs.
Investors are now purchasing geospatial technology, once used to monitor adversaries’ missile launch sites, to count cars in Walmart parking lots and to monitor crop yields in order to gain a competitive edge in investment decision-making. Imagery-focused commercial enterprises have found innovative ways to explain to new customers how these technologies can affect their bottom line. Moreover, these enterprises are now fusing satellite imagery with other data sources such as social media, allowing users to assess business risks (e.g., geopolitical conflicts, energy resources, natural disasters) or to obtain data feeds organized by location (e.g., ports, pipelines, borders).
This new business model presents several challenges for government. First, the technological growth and innovation in components of the commercial imagery sector, such as AI, may significantly outpace that of the national security arena. Much AI innovation will be focused, funded, and developed primarily for highly profitable commercial uses. Despite successful examples of public-private partnerships, national security uses of a company’s AI products may be a secondary objective for a growing number of businesses engaged in geospatial production and analysis. Second, the government is at risk of losing the battle for top AI talent. The private sector can pay more and offer more attractive and flexible workplaces than the government. Third, attempts to secure, classify, or restrict emerging AI technologies on national security grounds will be met with stiff resistance by the commercial sector, which has made significant research and development investments in anticipation of significant monetary returns. Another end result of these developments is that it will be easier for foreign governments to work with and acquire new technologies.
An implication of high-quality and affordable GEOINT is that NGOs, and an increasing number of U.S. and foreign government entities at all levels, are now able to use compelling imagery to put forward plausible analysis and interpretations about world events. In turn, these analyses and perspectives are easily broadcast via the internet and can reach ever-growing audiences at negligible cost. This new capability results in competing narratives with regard to developing security issues that must be sifted through and adjudicated by policy-makers worldwide. One example of how this phenomenon plays out can be seen in the various narratives offered up by different entities regarding territorial claims in the South China Sea.
The Growing Role of NGOs
As PTMs evolve and grow, NGOs will increasingly influence policy conversations leading to verification determinations. An example of this new phenomenon is the work of the James Martin Center for Nonproliferation Studies at the Middlebury Institute of International Studies at Monterey. The center’s analysis of and publications on its discovery of a North Korean missile production facility received wide dissemination both through traditional and social media outlets. The center has published several studies that have influenced the policy debate and expanded general public awareness about proliferation.
Despite the improvements in imagery analysis and interpretation, the techniques employed by NGOs may lead to faulty analysis and misinterpretation. Using overhead imagery effectively requires not only specialized supporting software but also geospatial expertise. The complex nature of working with satellite data and the lack of standardization of data is a challenge even for skilled imagery analysts. It is essential that users have the expertise to analyze the imagery and interpret the data, which includes choosing the best type and resolution of imagery for the intended illustration, along with the need to properly process the raw data. These requirements pose serious obstacles to NGOs that are often operating on a shoestring budget but eager to embrace the opportunities that geospatial information, data analytics, and social media present. Another challenge for NGOs is navigating complex commercial licensing arrangements that may limit public distribution of images due to restrictions found in government contracts with industry. On a positive note, some NGOs have arranged for discounted pricing on imagery used for non-commercial purposes. Nonetheless, the biggest risk to NGOs can come from publicizing faulty conclusions that can potentially tarnish their credibility and the reputations of those relying on their analysis.
As NGOs establish expertise in the use of GEOINT, they are increasingly organizing their approach to information using methods similar to those of government intelligence services. NGOs face many of the same policy decisions about information access, control, and influence as government intelligence services, and must decide when and how best to disseminate satellite imagery and geospatial analysis on a particular policy question or situation.
Although most nonproliferation-focused institutions rely on basic electro-optical imagery, some are beginning to make wider use of radar, infrared, other spectral imaging, and advanced processing techniques. A decent proxy for capacity among these institutions is whether they can process imagery in-house, or whether they must rely on others to process the image. Groups can draw conclusions from visually examining images processed by others, but the ability to conduct in-house processing offers significant advantages. Some institutions working in this area include:
- 38 North (focuses on North Korea)
- AllSource Analysis
- The Atlantic Council’s Digital Forensic Research Lab
- The Center for Strategic and International Studies (CSIS) (“Beyond Parallel” focuses on North Korea)
- The Institute for Science and International Security
- The James Martin Center for Nonproliferation Studies, Middlebury Institute of International Studies at Monterey
- The Verification Research, Training and Information Centre
Monitoring and Verification
The U.S. government is beginning to explore ways of collaborating with these institutions. One such partnership is between CSIS and NGA. This collaboration reflects the realization that NGOs and the U.S. government can do a better job of analyzing threats in cooperation with each other as opposed to proceeding independently, and is driven by the Intelligence Community’s (IC) goal to provide greater transparency that enhances public understanding and promotes collaboration with those outside the IC. Moreover, if NGOs like CSIS can conduct expert analysis without using classified data, their analysis can be more easily shared with allies and even adversaries.
In addition, international organizations like the International Atomic Energy Agency (IAEA) and the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO) see value in the use of GEOINT to support their work. IAEA has an in-house capacity to use satellite imagery and referenced this resource together with other open sources in its most recent report on North Korea. Several NGOs have worked with IAEA to build the agency’s capacity in this area. CTBTO has a more restrictive mandate but is currently exploring how to build greater capacity in this area to support on-site inspection. Adding to seismic and other data collected and analyzed by CTBTO, commercial satellite imagery provides the precise geolocation of underground nuclear tests. Analysts at 38 North have used this capability in their work monitoring North Korea’s Punggye-ri Nuclear Test Site.
Ultimately, the impact of technology acquisition in the commercial sector and the use by NGOs of more complex and powerful tools will yield a more robust offering of geospatial and related data analysis, fueling international debate. This debate may be amplified by NGOs, which regularly publicize their findings in traditional media as well as through blogs and websites. The ability to appropriately evaluate a multitude of claims and counterclaims will be more difficult in the future, as there may be much more noise (all data and analysis) than there is signal (relevant and reliable analysis). In short, the job of GEOINT professionals working with and in government will become more difficult as the private sector and NGOs increasingly offer analysis of their own and compete for the dominant narrative. The days of government monopoly of the monitoring process are waning.
Still, the emerging capabilities of NGOs, as significant as they are, do not diminish the primacy of governments in monitoring and, more importantly, making verification determinations of compliance and national interest. These are policy judgments that can only be performed by the state parties to international agreements. Governments have a far greater capacity—largely through intelligence sources and methods as well as negotiated inspection, information-sharing, and confidence building measures—to discover and penetrate weapons programs of concern. Governments are well positioned to facilitate the participation of international organizations and NGOs in the monitoring process and to assess the credibility of their contribution to verification judgments.
The increasing influence of geospatial and other open-source information to the monitoring and verification process will pose challenges as well as opportunities for governments. Although such open-source information provides valuable data, it may introduce spurious information that complicates efforts to penetrate denial and deception in monitoring. It also creates opportunities for the manipulation of the policy process of verification. The verification process is increasingly taking place in a political environment in which suspicions, “fake news,” disinformation, and unfounded accusations flourish. The geospatial and nonproliferation communities will need to cooperate more closely to produce and authenticate data that must be seen as objectively unbiased and impartial, which is the lifeblood of effective verification decision-making.
- In this article, the term “monitoring” refers to the gathering of information relevant to compliance assessments, including imagery and other forms of remote sensing tools. “Verification” refers to the process of reaching policy judgments about the extent and significance of compliance or noncompliance.
- Commercial GEOINT Strategy – 2018 Update. National Geospatial-Intelligence Agency. https://www.nga.mil/Partners/Pages/Commercial-GEOINT-Strategy.aspx. Accessed August 20, 2018.
- Artificial intelligence (AI) is usually defined as the science of making computers do things that require intelligence when done by humans. It is an evolving nascent technology. Machine learning is a sub-set of AI that involves algorithms that can learn to make predictions over time without being explicitly programmed to do so.
- Christopher Stubbs and Sidney Drell. “Public Domain Treaty Compliance Verification in the Digital Age,” IEEE Technology and Society Magazine, Winter 2013.
- Richard Engel and Kennett Werner. “Open-Source Material Offers Hints on North Korea’s Missile Capabilities.” NBC News. March 1, 2018. https://www.nbcnews.com/news/north-korea/open-source-material-offers-hints-north-korea-s-missile-capabilities-n850246.
- See for example: Ellen Nakashima and Joby Warrick. “U.S. Spy Agencies: North Korea Is Working on New Missiles,” Washington Post, July 30, 2018. https://www.washingtonpost.com/world/national-security/us-spy-agencies-north-korea-is-working-on-new-missiles/2018/07/30/b3542696-940d-11e8-a679-b09212fb69c2_story.html?noredirect=on&utm_term=.c0311e1b4e0c; Jeffrey Lewis and Dave Schmerler. “North Korea Expanding Key Missile Site,” Arms Control Wonk. July 2, 2018. https://www.armscontrolwonk.com/archive/1205558/north-korea-expanding-key-missile-site/. Accessed August 30, 2018; Catherine Dill. “Open Silos,” Arms Control Wonk, August 22, 2018. https://www.armscontrolwonk.com/archive/1205826/open-silos/. Accessed August 30, 2018.
- CSIS Korea Chair Announces Research Partnership with National Geospatial-Intelligence Agency (NGA). May 22, 2018. CSIS Press Release. https://www.csis.org/news/csis-korea-chair-announces-research-partnership-national-geospatial-intelligence-agency-nga. Accessed August 20, 2018.
- Intelligence Community Directive 107. Office of the Director of National Intelligence. https://www.dni.gov/files/documents/ICD/ICD-107.pdf. Accessed August 30, 2018.
- Application of Safeguards in the Democratic People’s Republic of Korea. International Atomic Energy Agency. August 20, 2018. https://www-legacy.iaea.org/About/Policy/GC/GC62/GC62Documents/English/gc62-12_en.pdf.
- See for example: The Punggye-ri Nuclear Test Site: A Test Tunnel Tutorial. 38 North. May 23, 2018. https://www.38north.org/2018/05/punggyetunnel052318/.