Societal Verification of Nuclear Disarmament in the 21^st Century: A Workshop Report

ABSTRACT

The idea that citizens, including scientists, and civil society groups be responsible for the monitoring and verification of their state’s compliance with nuclear weapons reduction and prohibition treaties – complementary to state-level and international agency mechanisms – has had various iterations. Each iteration reflected a particular historical moment of nuclear politics and identities, state-society relations, and possibilities for communication and collective action. A 2023 Princeton University Program on Science and Global Security workshop explored contemporary civil-society practices related to nuclear program monitoring and verification given the prevalence of internet connectivity, online publicly available data and tools, and cheap ubiquitous sensing from mobile cameras to commercial satellites. The aim of the workshop was to assess the potential and limits of such societal verification efforts. The workshop had three broad themes: the present context of societies, technologies, markets, and states within which civil-society does nuclear-activity monitoring and verification; how the uneven, hierarchical national and international distribution of power and relevant resources is enabling, shaping, and limiting the potential contributions of civil society to national and global security debates and decision-making; and the factors underlying the trustworthiness of civil-society nuclear-activity monitoring and verification practices and their outcomes as well as the relationship between trust (in data, processes, and institutions) and balanced practices.

KEYWORDS:

• Societal verification
• nuclear monitoring and verification
• nuclear disarmament
• nuclear proliferation
• nuclear arms control
• Treaty on the Prohibition of Nuclear Weapons (TPNW)

Introduction

The idea that civil society could and should play a role in the monitoring and verification of nuclear weapons treaties for nuclear arms control, nonproliferation, or disarmament was laid out in 1946 as part of the post-Manhattan Project scientists’ movement by scientist and activist Leo Szilard in his essay “Can We Avert an Arms Race by an Inspection System?” (Szilard 1946). The idea continued to be of interest in the 1950s and 1960s (cf. Bohn 1961; Melman 1958, Clark and Sohn 1958; Szilard 1961), but this approach eventually was overwhelmed first by US Cold War concerns over spying and less visibly by the development in secret and monopolization by states of remote monitoring and verification technologies, especially high-altitude surveillance and airborne sampling planes and later imaging satellites able to monitor large-scale nuclear weapon facilities and activities. In 1992, Manhattan Project physicist and Pugwash movement co-founder Joseph Rotblat updated this proposal and introduced the term “societal verification” (Rotblat 1992), reflecting the hopes mobilized by the end of the Cold War and collapse of the Soviet Union and the prospect of sweeping action towards nuclear disarmament.

For Rotblat, one of the main obstacles to verifiable disarmament was trust in states’ declarations of their nuclear activities. Clandestine activity couldn’t be ruled out, given the role of secrecy in states’ military planning and capabilities, nor effectively verified through the circumscribed nonproliferation safeguards framework of the International Atomic Energy Agency (IAEA). This concern was made evident at that time by the discovery of undeclared nuclear activities in Iraq that had been going on for years, mostly outside the scope of the standard operating practices of IAEA safeguards inspections. The problem was imagined to be even more acute in the case of assuring formerly nuclear-armed states were meeting their disarmament obligations once they had agreed to such a treaty.

Rotblat argued that the solution to this problem cannot only be technical, echoing previous analysis. The paradigm of relying on an IAEA system of physical inspections, agreed measuring instruments and methods, and space-based and other overhead surveillance controlled by other states might well be inadequate, no matter how far those methods improved. Rotblat’s idea of societal verification envisaged a regime of international treaty-stipulated duties and treaty-sanctioned rights for citizens to report their own state’s treaty violations as a way to supplement and overcome the deficiencies of state-level and international verification approaches. Rotblat, like earlier advocates of the idea, recognized that in this new paradigm for monitoring and verification of nuclear activity there would need to be protections for citizens reporting any violations as well as an expansion of civil society identification and responsibility beyond the nation-state to an international citizenship and accountability. Societal verification thus aimed to furnish part of the political infrastructure of the disarmament verification process, alongside the narrowly technical aspects, with the two processes reinforcing each other.

Rotblat conceived of two models for civil society/citizen reporting, distinguished by the degree of participation in the operations of a given state’s nuclear program. The first model is that of whistleblowing by insiders in relevant facilities, industries, and disciplines, including but not limited to watchdog organizations of scientists and technologists. Insiders could monitor, for example, personnel recruitment, material and equipment procurement, the nature of scientific work being carried out, and processes of uranium enrichment, spent nuclear fuel processing, and material storage. The second model is that of civil society/citizen reporting by those on the outside of the complex. Such reporting could enable monitoring nuclear facility construction and operations as well as nuclear material transportation and storage through impacts on the surrounding environment and communities. During the negotiation of the only nuclear disarmament treaty to date, the Treaty on the Prohibition of Nuclear Weapons, there was a proposal to include a Rotblat clause to prescribe societal verification mechanisms and protections (Mian, Patton, and Glaser 2017a, 2017b). It was not agreed.

In the three decades since Rotblat’s suggestion of societal verification, the nuclear nonproliferation and arms control field has seen a flourishing in a handful of mostly Western countries of non-governmental organizations (NGOs) focused on tasks that bear a family resemblance to what he envisioned. Academic institutions and non-governmental and commercial research groups have been analyzing commercial or free satellite imagery and material on the internet using publicly available tools to infer certain kinds of nuclear activity considered “suspicious” and then sharing their analyses through news media and online to elicit some kinds of policy and public responses. An archipelago of such groups has been described as a non-state open-source intelligence (OSINT) community whose goal is to collect and deploy publicly available information to support requirements for national security decisions. The term OSINT explicitly mirrors its US intelligence agency counterpart and other terms in the US intelligence agency lexicon, including SIGINT (signals intelligence), HUMINT (human intelligence), and IMINT (imagery intelligence), among others. For OSINT is not only conducted by NGOs and independent analysts but is also part of routine intelligence collection by government agencies.

This development has been made possible in large part by technologies that enable civil society groups and individual citizens to expand their capabilities for monitoring, investigation, and reporting. It also has been a result of greater interest in some governments and some parts of the media in this kind of public-sourced information and analysis. Understanding the possibilities and limits of non-governmental institutions and practices in creating and distributing certain kinds of public knowledge relevant to nuclear monitoring and verification needs to recognize how they reflect, contest, and shape relations of power and understandings of Self and Other in the current order and like all power-knowledge systems bear “the stamp of our age and geography” (Foucault 1966).

In order to explore and assess the contemporary challenges, opportunities, and prospects for societal verification, Princeton University’s Program on Science and Global Security organized a workshop in Spring 2023. This report provides an overview of the topics that emerged in the discussions and reflections by the authors who also were the organizers of the meeting, but it is not intended to reflect a consensus view of workshop participants. Previous work by the authors on the topic of societal verification includes Al-Sayed (2022) and IPFM (2009).

Workshop Description

The one and a half-day exploratory international workshop “Societal Verification: Realizing Joseph Rotblat’s Vision in the Age of Non-state Open-Source Intelligence” brought together scholars, civil society and government experts, and participants from the United Nations (UN) and a commercial imaging satellite company, among others. The workshop covered a range of issues and the discussions are presented here in a structure organized around three major themes.

Societal Verification in Context: States, Societies, and Technologies

The first theme of the workshop was the different perspectives on the current context and diverse roles of civil-society actors in nuclear nonproliferation and arms control monitoring and verification and how these are shaped by key institutional and organizational contexts comprising states, societies, and technologies.

It was the need to challenge an intensifying Cold War nuclear confrontation and arms race in the 1980s that drove the emergence of a diverse array of increasingly expert and capable civil society analysts based in universities and research and activist groups focused on transparency and accountability that established a basis for what exists today as societal verification goals and practices. An early example is the Nuclear Weapons Databook Series, compiled by the Natural Resources Defense Council, based in Washington, D.C., that provided unprecedented independent open-source information in U.S. Nuclear Forces and Capabilities (Cochran, Arkin, and Hoenig 1984), U.S. Nuclear Warhead Production (Cochran et al. 1987), U.S. Nuclear Warhead Facility Profiles (Cochran et al. 1987), Soviet Nuclear Weapons (Cochran et al. 1989), and British, French, and Chinese Nuclear Weapons (Norris, Burrows, and Fieldhouse 1994).

These approaches gained strength and new tools with the relative easing of nuclear weapon secrecy constraints in both the United States and former Soviet Union in the wake of the end of the Cold War. Commercial agents in both the United States and Russia began selling satellite imagery from US and Russian spy satellites in the early 1990s. In 1995, previously classified US spy satellite imagery from the Corona program covering the 1960s and 1970s was released. In the 1990s, the Federation of American Scientists’ Public Eye program, directed by John Pike, pioneered the use of declassified imagery and the newly available commercial imagery to inform public discourse on military arsenals around the world and influence policy debates. Another factor was the emergence of the internet and commercial websites providing aerial and satellite imagery in the late 1990s, notably Microsoft’s TerraServer website. An early application of this emerging suite of sources and capabilities was a Princeton study of the size of the Soviet/Russian arsenal, locations and capacity of warhead storage sites, and warhead dismantlement rates, to assess the scope for further deep arsenal reductions through US–Russian arms control agreements (Handler 1999). As the superpower Cold War dynamic ebbed, and arsenals slowly if unevenly were reduced, analysts’ attention turned increasingly from arms reduction and disarmament by nuclear-armed states to nuclear proliferation crises.

A lack of trust in governmental processes in the early 2000s in light of unfounded US and UK allegations against Iraq of the illicit pursuit and possible possession of nuclear weapons and other weapons of mass destruction further motivated non-government analysts to pursue open-source research. These analysts sought to empower civil society to contest and shape official policymaking through robust public discussions grounded in expert analysis. This reflected an enduring legitimacy crisis over national security decision-making in the nuclear age and was part of longstanding struggles to hold governments accountable on decisions that had possible international and global impacts.

The work of many non-governmental analysts shares commonalities with academic research and journalism, especially with regard to a pursuit of objectivity of analysis and transparency of evidence and method. It differs in varying degrees from open-source analysis by intelligence agencies in terms of goals, capabilities, approach, and epistemic and moral values. However, such analysts are mostly based in the United States and Western Europe and face political challenges in exporting to other societies the model of agency and transparency for the demos by the demos. Their interventions also trigger pushback from international bodies charged with monitoring and verification by nonproliferation and arms control treaties, including the IAEA and the Comprehensive Nuclear-Test-Ban Treaty Organization.

Satellite imagery has come to play an increasingly visible role in non-governmental monitoring and verification analysis and interventions, with a picture still seeming to offer irrefutable proof where text is seen as presenting a contestable argument. Commercial satellite imagery providers are now the dominant source of such information for analysts to process and give meaning. These companies commonly face challenges in making tradeoffs between public service provisions (for instance making their services accessible to all, reaching out to marginalized groups, etc.) and the market logic of achieving profit goals. Commercial providers typically manage this tradeoff through a vetting process for customers based on knowledge, skills, and accountability for the responsible use of services. This creates formal and informal clusters of imagery providers and clients. Governments are often the largest purchasers of imagery.

The UN bodies rely on open sources for parts of their work, where the analysis informs internal processes, such as in the case of the IAEA for the verification of safeguards agreements. The analysis may also be conducted in response to requests for clarity by the international community on events of geopolitical import. There are associated merits to and concerns with the UN’s work with open sources from the perspectives of the UN itself and states parties, especially in regard to the question of the neutrality of the sources. That being said, the United Nations Office for Outer Space Affairs (UNOOSA) conducts capacity building particularly for developing countries in the use of space technologies, including satellite technology, for sustainable social and economic development. Education and training in the use of the same resources for monitoring in the domain of international security isn’t, however, part of UNOOSA’s mandate. The separation reflects the compartmentalization of UN work and echoes the bracketing of human security concerns from international security.

Some governments provide the imagery obtained through their Earth-observation (non-military) satellites for free for citizens from around the world. The US government has done this on a large scale, especially with the Landsat satellite program launched in the 1970s, but also with the decision in 1995 to publicly release older imagery acquired by reconnaissance satellites, including from the Corona program (1959–1972). Another example is Brazil, including through international cooperation initiatives in space-related activities in the BRICS context to address through Earth-observation satellites challenges related to global climate change, disaster management, environmental protection, and food and water scarcity prevention. Brazil also acts through the UN on space policy and capacity building as well as through other international platforms for the coordination of space activities and promotion of data availability, access, or use for social and economic benefits. The Brazilian example offers insight into space-related capabilities and perspectives available in some non-nuclear weapon states in the Global South, and taken together with Brazil’s membership in the Latin America Nuclear Weapons Free Zone Treaty and the Treaty on the Prohibition of Nuclear Weapons offers a prospective broadening and democratization of nuclear-activity monitoring and verification. This could be explored in a regional conference to stake out opportunities for societal verification from a regional lens, taking into account technical, political, legal, and ethical perspectives.

Power, Ways of Seeing, and Voice

The second theme concerned the role of the national and international distribution of resources necessary for civil-society monitoring and verification practices in enabling, shaping, and limiting the potential contributions of civil society to global security debates and decision-making. Obviously necessary resources are those of the technical and financial kind, basic to the process of data acquisition and analysis for nuclear-activity monitoring. But the media and politics also determine the scope of academic and think-tank open-source analysis and policy intervention.

The distribution of resources that enable societal verification currently favors the United States and to a lesser extent Western Europe, simply reflecting existing social, political, and technical processes and institutions. This concentration of capabilities also serves, however, to inflect the orientation of much civil-society effort preferentially to sites and activities in countries already the focus of concern by political leadership in the United States and its allies. This “selective gaze” (Rothe and Shim 2018) is helped by and reinforces the “clandestineness” narrative (Lawrence 2020) – if it is being hidden it is suspicious, and a threat, while transparency is evidence of innocence and the key to security.

Non-governmental monitoring practices built on this clandestineness narrative may contribute to fueling security dilemmas, exacerbate existing inequalities, and create new concerns. The practice in simple generic terms can be described as cycles starting with national policies about some adversary countries and their programs (or a more specific tip in some cases), leading analysts to the acquisition of images from commercial satellite providers, then the analysis of this imagery and other open sources to draw conclusions, then sharing the analysis with public media channels, which in turn triggers discussions in national and international political fora and the launch of calls for investigation. Any investigation can then prompt a new cycle of activity culminating in more questions. The process affirms repeatedly the clandestineness narrative.

Beyond helping constitute the object of the selective gaze from above – a vantage point cast in modern intellectual history as objective by virtue of (equi)distance – and more broadly helping reshape geographical and geopolitical imaginations, satellite imagery throws up a host of questions related to representation and knowledge production. Art critic John Berger once observed, “The relation between what we see and what we know is never settled […] The way we see things is affected by what we know or what we believe […] We only see what we look at. To look is an act of choice […] Every image embodies a way of seeing” (Berger 1972). Building on this, what are the factors influencing choices made regarding where to look for evidence, say, of wrongdoing? How is knowledge derived from the interplay of the expectation of wrongdoing and its representation? What can be inferred by what we (want to) see and how we (are made to) see (Rothe and Shim 2018)? What remains obscured and how is inherent obscurity instrumentalized by the observed party? More generally, how does one evaluate claims to knowledge, authority, and responsibility in view of transformative ways of “looking”, using new media – satellite imagery or otherwise? What is the role of academic, journalistic, legal, and ethical standards in ensuring a level of objectivity and fairness in representation?

The inevitable result and cost of selective attention is “crowding out” patterns. For the imagery provider, those patterns manifest in the scalability of their business being a function of demand, in customer vetting based on skills (privileging seasoned, data-intensive users, with the agriculture and insurance industries being large markets) and intentions, in founder political preferences and their aptitude for social benefit end uses, and in inundating prospective clients with pricing structure navigation and service picking-and-choosing. It remains unclear how to engage commercial stakeholders in a transnational pursuit of global equity in view of the entrenchment of a technical–managerial logic where means predominantly dictate ends. On the other hand, the UN enjoys little agency and is already logistically challenged. It is reactive in attempting to bring clarity and authoritative assessment to news-cycle driven issues based on second-hand publicly available information, including third-party analyses that may follow divergent interests to those of states parties.

Many in the non-governmental open-source monitoring and verification community entertain the belief that transparency is the new and unavoidable reality and norm for governments and societies, and endorse the adage, “knowledge is power” – power for all publics and all policymakers. This universally inclusive framing may, however, gloss over and reproduce actually existing technopolitical arrangements that determine the loci of knowledge production, and power, precluding rather than affording agency to others. Because it is seen as lopsided and coming from power centers, the push for transparency and accountability may undermine trust within already marginalized parts of the international community. At the same time, the notion that disclosing nuclear information is always good may be seen to conflict with the perspective that transparency may drive proliferation and arms racing. Finally, limiting the circulation of knowledge to allow power to do its work – for instance, by withholding information for the sake of securing private diplomacy – entails sacrificing fundamental democratic values and rights for a guardianship model of governance (Dahl 1985).

Insofar as more data is seen to be the route to empowerment and the democratization of the nuclear-activity monitoring and verification sphere, the lack of digital literacy is also a barrier to democratized data access. This poses the question: what technical, political, financial, legal, and ethical resources need to be put in place to set up practices of contestation using open sources, when the influx of potentially more data and subsequently more analysts must engage with a status quo narrative and a paradigm that determines what a problem is and what a solution looks like.

A final related question is whether today’s nuclear non-state OSINT community shares a clear common purpose. It probably does not. Recent discussions over ethics in OSINT practice suggest there is still a fundamental institutional gap to be bridged, should ethics become anything more than an abstract construct with seemingly arbitrary modes of application. (For example, the Stanley Center for Peace and Security launched an initiative in 2019 in cooperation with stakeholders looking at ethical issues in the nuclear non-state OSINT community through a series of activities. A gist of the activities could be gleaned from the reports by the Stanley Center for Peace and Security and Open Nuclear Network (2020) and Loehrke et al. (2022).)

But it is also not clear whether it is necessary or feasible to further institutionalize this community – in view of the emerging assemblage of processes, institutions, and corporations – into a profession and regulate it, let alone organize it into manifesting a common purpose. For another perspective holds that maintaining the current level of diversity is healthier, so as not to foreclose serendipitous discoveries resulting from different practices, motivations, end uses, cultures, worldviews, etc. With time, institutions would perhaps arise, and with those, trust would be cultivated. Another concern shared by skeptics of institutionalization is the likelihood this endeavor would mar the community with the tension between expertise and democracy and possibly narrow the community’s range of actions to the detriment of the cause of verification. On the other hand, institutionalization might increase the chances of state funding, where funding is direly needed to improve and expand societal-verification capacities.

Trust, Balance, Accountability, and Contentious Politics

The third workshop theme was the elements that make up the trustworthiness of civil-society nuclear-activity monitoring and verification practices and their outcomes. This includes the relationship between trust (in data, processes, and institutions) and balanced practices.

The overlaying of Rotblat’s conception of societal verification and its contemporary manifestations begs two questions. The first is the question of loyalties and obligations to their state of individual citizens and of citizen groups organized for collective action as civil society versus that of a human being to humankind, transcending the prevailing structure of nation-states. The second question is about the emancipatory potential of societal monitoring practices (in the nuclear weapons domain or otherwise) and the possibilities they offer in resisting structures of domination and violence, including the state and the system of states.

The latter question turns in part on the nature of the right to knowledge in relation to prima facie duties of beneficence and nonmaleficence – specifically, when it is a matter of harm and violence, who should know what, about what, from whom, according to whom, and to what end? Relevant here is the practice of civil-society monitoring of human rights violations in many democratic countries and following the signature of the Helsinki Final Act in 1975 in the Soviet Union and its Warsaw Pact allies. It is noteworthy that in the human rights and humanitarian action fields the use of open sources is gaining ground faster in use and legitimacy compared to the monitoring of nuclear weapons programs. The factors driving this development are twofold: the growing salience of guidelines for improving the quality of sources used in international criminal, human rights, or humanitarian investigations and the emerging culture of best-practice sharing and establishment of core principles, while accounting for the different evidentiary standards depending on the end use of analysis, whether journalistic or legal.

The experience of the human rights and humanitarian action communities raises the question of how societal verification concerning nuclear weapons, including the practices of the nuclear non-state OSINT community, can create and turn data into evidence in the absence of international laws on a par with international criminal, human rights, and humanitarian law. What would the relevant processes look like, beyond publishing analyses in the open media?

There are parallels with and lessons to be learned from other contexts such as NGO environmental and human rights open-source investigation work. One instance is that of cases of ecocide and environmental injustice in the absence of an adequate legal framework for prosecution and in view of prevailing partial and contested notions of environmental crimes. Another is how data become evidence in cases of harm from nuclear weapons testing and the struggles with and within governments over investigation results and the adequacy of reparative actions.

Efforts should be invested to translate the lessons learned from those contexts to the monitoring and verification of nuclear nonproliferation, arms control, and disarmament obligations. For example, if legal structures are key to establishing the means for data to become evidence, an effort could be launched to create a formal legal basis for the prosecution of any and all states for the development, testing, production, stockpiling, stationing, transfer, use and threat of use of nuclear weapons, as well as any assistance to those activities. In this regard, could the Rome Statute of the International Criminal Court be amended to criminalize nuclear weapons? The Court has responsibility for investigating and punishing crimes against humanity, crimes of genocide, and war crimes, and in 2009 Mexico officially proposed that the use or threat of use of nuclear weapons be listed as a war crime under Article 8 of the Rome Statute (ICC 2009).

While operation within an established legal framework goes a long way to promoting trust, another means to promoting trust is granting agency to the communities directly impacted by the nuclear weapons complex, from the mining of nuclear material to nuclear waste disposal. It is pivotal to trust-building to involve these communities in nuclear-activity monitoring practices, so they have ownership over the process and outcomes. For a start, those communities could be engaged in speculative exercises where a nuclear-weapons–free future in their particular contexts is to be imagined along with possible pathways to achieving it. This could include the role of societal verification and its operationalization – i.e. what indicators for disarmament to look out for and how to monitor them – while accounting for constraints (including legal and ethical) unique to each context. In addition to political mobilization at the grassroots level, such an effort could pave the way to mitigating the imbalance in the current monitoring environment.

Trust concerns may become amplified to the extent that machine-learning(ML)–based approaches would inevitably have to be leveraged to analyze the increasingly large volumes of publicly available information with bearing on treaty verification. For example, ML-based approaches could be used to analyze satellite imagery, conduct transit matching, and review science and technology publications. Challenges impacting trust include insufficient training data of relevance to nonproliferation, arms control, and disarmament, in addition to the lack of high-quality annotations – if annotations are available to begin with. Nevertheless, considerations around the incentivization and vetting of participants as well as data and labor protection should be factored in. Moreover, the general impacts of reliance on ML-based verification systems on trust in current and prospective future treaty regimes should be carefully examined in view of non-trivial challenges such as data bias and adversarial data poisoning, limitations to model explainability, and privacy and other ethical, legal, and political concerns.

A merit to involving the public in data collection and annotation and model-building efforts is that this could be a pathway to raising the public’s awareness of nuclear issues and building community with a sense of agency in nuclear matters that de facto and de jure are the exclusive province of governmental and intergovernmental actors. Deliberate and deliberative expert practices aimed at involving communities in preparing for and engaging in societal verification may help usher in the “change of attitudes” underpinning Rotblat’s vision.

For an advocate of Rotblat’s vision, an essential attribute of the idealized societal-verification community is its politics of contention of citizens against their governments. The community’s purpose would be that of citizens holding their governments to account under domestically ratified international law banning nuclear weapons. In the absence of this purpose, societal verification practices may promote mistrust or distrust among publics and governments, with destabilizing effects.

Acknowledgments

Special thanks go to Sarina Hegli, Christopher Lidard, and Gigi Schadrack for taking thorough notes that formed the basis of work on this report.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Notes on contributors

Sara Al-Sayed

Sara Al-Sayed is a Postdoctoral Research Associate in the Program on Science and Global Security at Princeton University. Her research in Princeton centers on how new and emerging information technologies can support data collection, analysis, and secure reporting to empower civil society with regard to nuclear arms control, non- proliferation, and disarmament verification.

Alexander Glaser

Alexander Glaser is an associate professor in the Princeton School of Public and International Affairs and in the Department of Mechanical and Aerospace Engineering at Princeton University. He is co-director of the Program on Science and Global Security.

Zia Mian

Zia Mian is a physicist and co-director of Princeton University’s Program on Science and Global Security, which is part of the Princeton School of Public and International Affairs.