Introduction
The international legal framework governing nuclear weapons rests on an unusual arrangement. The Treaty on the Non-Proliferation of Nuclear Weapons (NPT) lays down provisions to prevent the spread of nuclear weapons. This is done while recognising a distinction between states that possessed such weapons before a specified date and those prohibited from acquiring them. This distinction was accepted as part of a broader understanding that included nuclear disarmament for the parties that possessed the weapons and the peaceful use of nuclear energy.1 However, more than five decades after the Treaty entered into force, the continued reliance on nuclear weapons raises important questions about the implementation of this arrangement. According to the Stockholm International Peace Research Institute (SIPRI), nine nuclear-armed states possessed approximately 12,187 nuclear warheads as of January 2026. Of these, around 9,745 were held in military stockpiles, of which approximately 4,012 were deployed with operational forces. All nine states continued programmes to modernise or strengthen their nuclear arsenals during 2025.2 These figures do not establish a violation of international law by themselves. They nevertheless demonstrate the continuing importance of nuclear weapons in national security policies despite longstanding commitments towards disarmament.
The difficulties that the NPT is facing were also visible at its Eleventh Review Conference, held from 27 April to 22 May 2026. The Conference concluded without adopting a consensus substantive outcome, following the precedent of 2015 and 2022.3 The disagreements concerning Iran’s nuclear programme contributed to the final deadlock, alongside wider differences over disarmament and nuclear risks.4 Even the earlier commitments had a similar fate. The repeated failure to agree on a substantive outcome raises a question beyond the political divisions between states. How effectively can the NPT assess the implementation of its obligations when the review process itself depends heavily on political agreement?
Under Articles II and III of the NPT, the non-nuclear-weapon states undertake not to acquire nuclear weapons and accept safeguards designed to verify that relevant nuclear material is not diverted to prohibited purposes. Article VI, by contrast, requires all parties to pursue negotiations in good faith towards nuclear disarmament but does not establish an equivalent standing mechanism for reviewing the fulfilment of that obligation.5 This paper describes the resulting accountability problem as the “Nuclear Black Box”. The term refers to the difficulty of examining information about nuclear arsenals or disarmament negotiations (which is not completely unavailable) through consistent legal and institutional standards. States may disclose their policies, report reductions or participate in negotiations, yet the NPT lacks an agreed procedure capable of systematically assessing these developments against their Article VI commitments. The paper argues that the NPT’s principal vulnerability does not arise solely from its unequal allocation of obligations. Differentiated treaty obligations are not inherently unlawful. The more serious concern is the gap between the binding nature of disarmament commitments and the limited arrangements for assessing their implementation. The analysis proceeds through a doctrinal examination of the NPT, followed by relevant decisions of the International Court of Justice, the development of the review process, and recent proposals for institutional reform. It concludes by proposing a practical method for preserving and examining evidence of implementation even when states cannot agree on a collective political assessment.
The legal structure of the NPT
A. Differentiated obligations under the Treaty
The aim of the NPT was to prevent the spread of nuclear weapons while maintaining cooperation in the peaceful use of nuclear energy. It established two types of states: nuclear-weapon states and non-nuclear-weapon states. Article IX(3) of the NPT defines a nuclear-weapon state as one that manufactured and exploded a nuclear weapon or other nuclear explosive device before 1 January 1967.6 This definition recognises China, France, Russia, the United Kingdom, and the United States as nuclear-weapon states. Under Article I of the NPT, nuclear-weapon states undertake not to transfer nuclear weapons or to assist non-nuclear-weapon states in acquiring them. Article II places restrictions upon non-nuclear-weapon states by prohibiting them from receiving, manufacturing or acquiring nuclear weapons or other nuclear explosive devices.7 The Treaty does not require its recognised nuclear-weapon states to eliminate their arsenals immediately. Instead, Article VI establishes an obligation to negotiate on nuclear disarmament. The resulting framework is evidently unequal, as the states recognised as nuclear-weapon states retain weapons that other parties are prohibited from acquiring. However, states frequently accept different responsibilities under negotiated international agreements. The central question is whether the obligations undertaken by each category of states are implemented in good faith and receive meaningful institutional scrutiny.
B. The relationship between non-proliferation and disarmament
The NPT has three major objectives: preventing nuclear proliferation, facilitating peaceful nuclear cooperation, and advancing nuclear disarmament. These objectives are related and legally distinct. The restrictions imposed by Article II do not depend upon nuclear-weapon states first achieving disarmament. Similarly, the continued compliance of non-nuclear-weapon states does not remove the obligations established under Article VI. The NPT should not be understood as a simple contractual arrangement in which failure to fulfil one objective automatically releases other parties from their legal commitments. Under Article 31 of the Vienna Convention on the Law of Treaties, a treaty is interpreted in good faith in accordance with the ordinary meaning of its terms in their context and in the light of its object and purpose.8 Applying this approach, Article VI must be recognised as a binding obligation. However, its wording cannot be interpreted as imposing a particular deadline for the disarmament objective. This distinction is imperative because the credibility of the NPT depends upon preserving both its non-proliferation protections and its disarmament objectives. Weakening one objective cannot be justified by dissatisfaction with progress under the other.
Article VI and the requirement of good faith
A. The legal meaning of Article VI
Article VI requires each party to pursue negotiations in good faith on effective measures relating to the cessation of the nuclear arms race and nuclear disarmament. It also extends to a treaty on general and complete disarmament under strict and effective international control.9 The provision creates a binding legal obligation. However, it does not specify any numerical reduction targets or a fixed timetable. It does not even mention a supervisory institution responsible for assessing compliance. It is purely based on good faith, which is central to its interpretation. In practical terms, the obligation requires states to approach negotiations seriously and consistently with the purpose of nuclear disarmament. In order to establish relevant conduct, states may participate in negotiations, implement arms-control commitments, or showcase the advancement of credible proposals and measures intended to reduce nuclear risks. However, a finding of non-compliance would require an investigation of the applicable legal obligation and the conduct of the state concerned. The difficulty is that the NPT provides no common institutional method for making such assessments.
B. The International Court of Justice’s interpretation
The International Court of Justice examined nuclear disarmament obligations in its 1996 Advisory Opinion on the Legality of the Threat or Use of Nuclear Weapons. The Court identified that the disarmament obligation does not imply simply participating in diplomatic discussions. It unanimously affirmed an obligation to pursue in good faith and bring to a conclusion negotiations leading to nuclear disarmament in all its aspects under strict and effective international control.10 This interpretation is significant as it confirms that nuclear disarmament is not an aspirational political objective. States cannot treat participation in discussions as sufficient without regard to their purpose. The Advisory Opinion therefore strengthens the legal significance of disarmament commitments. However, it did not resolve the institutional question of how their implementation should be evaluated.
C. The limits of judicial enforcement
The difficulties associated with legal enforcement were illustrated by the proceedings brought by the Marshall Islands against several nuclear-armed states. In Obligations Concerning Negotiations Relating to Cessation of the Nuclear Arms Race and to Nuclear Disarmament (Marshall Islands v. United Kingdom), the Marshall Islands alleged failures related to nuclear disarmament negotiations. On 5 October 2016, the International Court of Justice upheld a preliminary objection as the existence of a dispute between the parties had not been established at the time the application was filed.11 The Court consequently did not decide whether the respondent had breached its disarmament obligations. This judgment does not undermine the legal potency of Article VI. It illustrates the procedural limitations that may prevent international courts from examining the substance of alleged violations. Judicial enforcement depends upon jurisdiction and other procedural requirements. It cannot by itself provide a continuous mechanism for assessing disarmament progress. The absence of such a mechanism makes the role of the NPT review process especially important.
The Nuclear Black Box: verification without equivalent disarmament accountability
A. The distinction between safeguards and disarmament assessment
The International Atomic Energy Agency (IAEA) verifies nuclear material through the safeguards agreements provided for by Article III of the NPT. These safeguards are intended to provide assurances that nuclear material is not being diverted from peaceful purposes to use in nuclear weapons or other nuclear explosive devices.12 However, the safeguards system has limitations. Its effectiveness depends upon the legal scope of the applicable agreements, available information, cooperation by states, and the Agency’s technical capabilities. The Additional Protocol, wherever in force, strengthens the IAEA’s access to information and locations; however, it is not automatically binding upon every NPT party.13 Still, the safeguards provide an established system of technical verification. Article VI does not contain a comparable standing mechanism for evaluating disarmament negotiations. This difference is partly justified by the nature of the respective obligations. Verifying nuclear material involves technical processes that cannot simply be applied to diplomatic conduct. Assessing good-faith negotiations requires consideration of proposals, policies, strategic circumstances, and state behaviour. The absence of identical verification methods is therefore understandable. However, the difference between these obligations does not justify the absence of consistent procedures for reviewing disarmament implementation. The central problem is not that Article VI cannot be verified in precisely the same manner as nuclear material. It is that the information relevant to Article VI is not examined through a sufficiently continuous and agreed institutional process.
B. Defining the Nuclear Black Box
The concept of the Nuclear Black Box refers to the gap between legally recognised disarmament obligations and the institutional capacity to evaluate their implementation. Three related difficulties create this gap. First, availability of information is not even across states. They disclose different quantities of information concerning nuclear arsenals, military doctrines, risk-reduction measures and disarmament negotiations.14 Some information is publicly available, while other information remains restricted for security reasons. Second, there is no uniform method of consistently evaluating the available information. In order to show progress, information such as a reduction in deployed warheads, participation in negotiations, or adoption of a risk-reduction measure may be a positive indication. However, the legal significance of such conduct under Article VI cannot be determined through a single numerical indicator. Third, there is a lack of sufficient institutional continuity in the review process. Information is submitted during different review cycles, but the failure to adopt agreed substantive outcomes can make it difficult to maintain a common record of progress and unresolved queries. Although the nuclear-armed states have, at different times, provided information about their policies and arsenals, and international organisations and independent research institutions also publish assessments, the problem is that available information does not automatically produce accountability. An effective review process must provide a means of organising evidence, evaluating its significance, and preserving unresolved disagreements over time. Therefore, the Nuclear Black Box is best understood as a problem of systematic and continuous institutional assessment.
The NPT review process
A. From the 1995 extension to successive review commitments
The 1995 NPT Review and Extension Conference marked an important stage in the Treaty’s development. Under Article X(2), states parties decided that the NPT would remain in force indefinitely. The decision was accompanied by arrangements intended to strengthen the review process and identify principles and objectives for nuclear non-proliferation and disarmament.15 The indefinite extension of the Treaty did not authorise indefinite possession of nuclear weapons without regard to Article VI. It preserved the NPT’s legal framework, including its disarmament obligations. Later Review Conferences attempted to give greater practical meaning to those commitments. The 2000 Review Conference adopted thirteen practical steps relating to nuclear disarmament.16 The 2010 Review Conference adopted an action plan addressing the Treaty’s three principal areas, which included measures concerning transparency and reporting as well.17 These developments demonstrate that states parties have repeatedly recognised the need for more detailed consideration of disarmament progress. However, Review Conference commitments do not automatically amend the NPT, and their legal significance must be assessed according to their wording, the circumstances of adoption, and the applicable rules of treaty interpretation.
B. Existing proposals for transparency and accountability
The need for stronger reporting and review procedures is not a new concern. During the preparatory process for the 2026 Review Conference, states advanced proposals for more regular reporting, structured discussion, and clearer indicators of disarmament progress. A working paper submitted during the preparatory process proposed improvements to Article VI monitoring, including national reports, structured discussion, and greater attention to the implementation of earlier commitments.18 In 2025, the United Nations Institute for Disarmament Research also examined possible models for interactive dialogue within the NPT review cycle, drawing comparisons with other international review mechanisms.19 Further proposals were advanced at the 2026 Review Conference, including a multistate working paper addressing transparency and accountability within the review process.20 These initiatives are important because they demonstrate that the problem of disarmament accountability is already recognised by states and researchers. They also place limits on claims of originality. Standardised reporting and interactive dialogue cannot be presented as entirely new solutions. The contribution of this paper lies in identifying an additional institutional problem: even improved reporting may be futile when the information provided is not consistently preserved and examined so that states can agree on substantive conclusions.
C. Lessons from the 2026 Review Conference
The 2026 Review Conference provides a recent illustration of this difficulty. Despite extensive negotiations, the Conference ended without agreement on a consensus substantive outcome. Disputes involving Iran’s nuclear programme and broader disagreements over disarmament and security contributed to the impasse.21 It would be incorrect to conclude that the Conference performed no useful function. States examined treaty implementation, expressed their positions, and discussed possible future measures. Article VIII(3) provides for reviewing the operation of the Treaty.22 It does not make the continuing legal validity of Article VI dependent upon the adoption of a consensus outcome document. The repeated absence of agreed substantive outcomes reveals a weakness in the system’s ability to produce a continuous collective assessment. Political disagreement may prevent states from adopting common conclusions. It should not prevent the preservation of factual information, national submissions, and competing interpretations. The 2026 experience demonstrates why procedural continuity should be treated as an objective distinct from political consensus.
Towards a continuous and evidence-based review framework
A. The proposed implementation record
The reforms proposed in this paper build upon existing discussions of transparency and structured dialogue. States should keep an Article VI Implementation Record within the existing Article VIII(3) review process rather than establish a new body. The purpose of this record would be to ensure better organisation of information and records of disarmament-related developments and to provide a systematic, structured account, including during periods of no consensus. The record could include four categories of information. First would be the national implementation reports. These would document measures reported by states, including relevant negotiations, arms-control agreements, nuclear risk-reduction initiatives, and implementation of earlier commitments. Second would be references to available supporting information. Where appropriate, reports could identify official documents, publicly available data, and other material relevant to the measures described. The third category would include unresolved questions and differing interpretations. States would be able to identify concerns about implementation without requiring all parties to endorse a common conclusion. Last would be information on progress across successive review cycles. Commitments and reported developments would be organised so that later meetings could examine whether previously identified issues had been addressed. The central advantage of this approach is continuity. An implementation record’s purpose would be to maintain an accessible and reliable account of information submitted during the review process.
B. Legal basis and institutional feasibility
The legal foundation for periodic review of the Treaty’s operation is provided by Article VIII(3). The review process was strengthened by the arrangements adopted in 1995.23 These provisions provide an institutional basis for considering improvements to reporting and documentation. However, it is necessary to distinguish between reforms that can be undertaken voluntarily and those requiring agreement among states parties. Individual states or groups of states may voluntarily submit reports in a common format, publish implementation information, or prepare comparative documents based on publicly available material. In contrast, establishing a new mandatory reporting obligation or assigning formal assessment powers to an institution would require an appropriate legal or institutional basis. Such authority cannot simply be assumed from Article VIII(3). A practical approach would therefore begin with voluntary reporting arrangements and the systematic use of existing conference documentation. The Preparatory Committee process could facilitate discussion of common reporting categories. States could also support a factual compilation of submissions and outstanding questions, subject to agreed procedures and available administrative resources. Such a record would not possess independent authority to determine legal violations. Its purpose would be to improve the quality and continuity of institutional review.
C. Measuring progress without creating artificial legal tests
A further challenge concerns the criteria used to evaluate disarmament implementation. Some developments are easier to measure than others. For example, changes in declared nuclear stockpiles, entry into force of arms-control agreements, and participation in negotiations can be documented easily. However, these factors do not automatically determine whether a state has fulfilled the obligations associated with Article VI. For this reason, the proposed review framework should distinguish between three levels of assessment. The first would focus on reported facts, which would include participation in negotiations or implementation of a particular measure. The second would underscore the significance attributed to those facts, including whether states consider them evidence of progress or continuing difficulty. The third would deal with the legal conclusions and would require interpretation of the applicable obligation and the circumstances of state conduct. In order to reduce the risk of political disagreement overshadowing legal conclusions, these levels would ensure that useful evidence is still preserved. They would also prevent indicators from being treated as substitutes for legal analysis.
D. Addressing security and sovereignty concerns
Nuclear-weapon states may resist extensive reporting on the ground that disclosure could affect national security. This concern cannot be dismissed. Information about nuclear forces may have strategic and operational consequences. Therefore, the proposed framework should be mindful that not every minute detail would be made public. States could distinguish between information suitable for public reporting and information they are unwilling or unable to disclose. The record should identify limitations in the available information. A second objection is that structured assessment may politicise the review process further. That risk does exist. However, political disagreement already forms part of the NPT review system. Separating factual documentation from collective legal or political conclusions may reduce, rather than increase, the extent to which disagreement prevents institutional progress. Finally, the proposed framework cannot compel states to disarm or guarantee meaningful negotiations. It is valuable because it would make the examination of implementation of existing commitments easier over time.
Conclusion
The Treaty on the Non-Proliferation of Nuclear Weapons remains central to the international legal framework governing nuclear security. Its differentiated obligations reflect a historical compromise intended to prevent the further spread of nuclear weapons while advancing peaceful nuclear cooperation and disarmament. The unequal nature of this arrangement has long attracted criticism. However, the existence of different obligations does not, by itself, establish a legal defect. The more important question is whether the Treaty’s obligations are accompanied by procedures capable of assessing their implementation. Article VI establishes a binding obligation to pursue negotiations in good faith towards nuclear disarmament. The International Court of Justice has confirmed the legal significance of this commitment. Yet the NPT does not provide a standing institutional mechanism comparable to the safeguards system for evaluating disarmament implementation. This paper has described the resulting problem as the Nuclear Black Box: the gap between legally recognised disarmament obligations and the institutional capacity to examine their fulfilment consistently. The failure of the 2026 Review Conference to adopt a consensus substantive outcome further demonstrates the limitations of a process in which collective assessment is closely tied to political agreement. Existing proposals for greater transparency and structured dialogue represent important steps towards addressing these difficulties. However, reporting alone cannot ensure accountability if information is not preserved and examined across successive review cycles. The proposed Article VI Implementation Record seeks to address this additional problem. In times when consensus is not available, the proposed Article VI Implementation Record would strengthen the review process by organising national reports, identifying unresolved disagreements, and maintaining continuity, without creating new substantive obligations or claiming authority to determine treaty violations. Such reforms are directed neither to curtailing strategic disagreements nor to guaranteeing compliance with Article VI. They are intended to strengthen the basic requirement of effective treaty governance by ensuring meaningful and continuous scrutiny of commitments. The long-term credibility of the NPT depends upon ensuring that the commitments made towards the eventual elimination of nuclear weapons can be assessed with greater consistency, transparency, and legal care.
*****
Footnotes
1. Treaty on the Non-Proliferation of Nuclear Weapons, opened for signature July 1, 1968, 21 U.S.T. 483, 729 U.N.T.S. 161 (entered into force Mar. 5, 1970) [hereinafter NPT].
2. Stockholm Int’l Peace Rsch. Inst., SIPRI Yearbook 2026: Armaments, Disarmament and International Security, Summary 10–11 (2026), https://www.sipri.org/sites/default/files/2026-06/yb26_summary_en_1.pdf.
3. U.N. Secretary-General, Statement Attributable to the Spokesperson for the Secretary-General on the Eleventh Review Conference of the Parties to the Treaty on the Non-Proliferation of Nuclear Weapons (NPT) (May 22, 2026), https://www.un.org/sg/en/content/sg/statements/2026-05-22/statement-attributable-the-spokesperson-for-the-secretary-general-the-eleventh-review-conference-of-the-parties-the-treaty-the-non-proliferation-of-nuclear-weapons-.
4. Review Conference Ends Without Consensus Outcome Amid Rising Nuclear Risks, U.N. Meetings Coverage, U.N. Doc. DC/3912 (May 22, 2026), https://press.un.org/en/2026/dc3912.doc.htm.
5. NPT, supra note 1, arts. II, III, VI.
6. Id. art. IX(3).
7. Id. arts. I, II.
8. Vienna Convention on the Law of Treaties art. 31(1), opened for signature May 23, 1969, 1155 U.N.T.S. 331 (entered into force Jan. 27, 1980).
9. NPT, supra note 1, art. VI.
10. Legality of the Threat or Use of Nuclear Weapons, Advisory Opinion, 1996 I.C.J. 226, 263–64, ¶¶ 98–103, 267, ¶ 105(2)(F) (July 8).
11. Obligations Concerning Negotiations Relating to Cessation of the Nuclear Arms Race and to Nuclear Disarmament (Marshall Islands v. United Kingdom), Preliminary Objections, Judgment, 2016 I.C.J. 833 (Oct. 5).
12. Int’l Atomic Energy Agency, IAEA Safeguards: Serving Nuclear Non-Proliferation (2024), https://www.iaea.org/sites/default/files/18/09/sg-serving-nuclear-non-proliferation.pdf.
13. Id. at 14.
14. Transparency and Improved Monitoring of the Implementation of Article VI, Working Paper Submitted by Ireland, New Zealand & Switzerland, Preparatory Comm. for the 2026 Review Conference of the Parties to the Treaty on the Non-Proliferation of Nuclear Weapons, U.N. Doc. NPT/CONF.2026/PC.1/WP.6, ¶ 2 (June 13, 2023), https://documents.un.org/doc/undoc/gen/n23/166/67/pdf/n2316667.pdf; Stockholm Int’l Peace Rsch. Inst., supra note 2, at 10.
15. 1995 Review & Extension Conference of the Parties to the Treaty on the Non-Proliferation of Nuclear Weapons, Final Document, NPT/CONF.1995/32 (Part I), annex (1995).
16. 2000 Review Conference of the Parties to the Treaty on the Non-Proliferation of Nuclear Weapons, Final Document, NPT/CONF.2000/28 (Parts I & II) (2000).
17. 2010 Review Conference of the Parties to the Treaty on the Non-Proliferation of Nuclear Weapons, Final Document, NPT/CONF.2010/50 (Vol. I) (2010).
18. Transparency and Improved Monitoring of the Implementation of Article VI, supra note 14.
19. James Revill, Victoria Viana Souza Guimarães & Luiza Delaflora Cassol, Exploring Reporting and Interactive Dialogue Options for NPT Disarmament Obligations 15–31 (U.N. Inst. for Disarmament Rsch. 2025), https://doi.org/10.37559/WMD/25/NDV/02.
20. A Path Towards Enhanced Transparency and Accountability Within the Review Process of the Treaty on the Non-Proliferation of Nuclear Weapons, Working Paper Submitted by Albania et al., U.N. Doc. NPT/CONF.2026/WP.1/Rev.1 (May 1, 2026), https://digitallibrary.un.org/record/4115438.
21. Review Conference Ends Without Consensus Outcome Amid Rising Nuclear Risks, supra note 4; Libby Flatoff & Daryl G. Kimball, 2026 NPT Review Conference Stymied by Disputes, Arms Control Today (June 2026), https://www.armscontrol.org/act/2026-06/news/2026-npt-review-conference-stymied-disputes.
22. NPT, supra note 1, art. VIII(3).
23. Id.; 1995 Review & Extension Conference of the Parties to the Treaty on the Non-Proliferation of Nuclear Weapons, Final Document, supra note 15, annex, Decision 1.