I. Introduction
On 29 January 2026, the International Court of Justice closed three weeks of merits hearings in The Gambia v. Myanmar and began its deliberations on whether Myanmar’s campaign against the Rohingya amounted to genocide under the 1948 Genocide Convention. This marks only the third occasion since the Convention entered into force that a state-to-state genocide claim has reached the merits stage before the Court, following the concluded proceedings in Bosnia and Herzegovina v. Serbia and Montenegro (2007) and Croatia v. Serbia (2015). Unlike those cases, the Court here lacks a resource that shaped both prior judgments considerably: a body of prior findings from an international criminal tribunal. No ad hoc tribunal has ever been established for Myanmar; the Court instead relies heavily on the United Nations Independent International Fact-Finding Mission (FFM) on Myanmar, whose own evidentiary threshold differs markedly from The Court’s.
This raises the question: can material gathered by the FFM under its own “reasonable grounds to conclude” threshold (UN Human Rights Council, 2018a, para.6.) satisfy, or otherwise be used to satisfy, the Court’s considerably stricter “only reasonable inference” test for genocidal intent under Article II of the Genocide Convention? At its core, this is a problem of evidentiary translation, whether material assembled for investigation can be repurposed to satisfy the different standard adjudication demands. The scope of this article is confined to that evidentiary question: it does not determine whether genocide occurred in Rakhine State, nor assess Myanmar’s international responsibility more broadly, since those matters remain for the Court’s pending judgment. Rather, it isolates a narrower problem that will recur wherever a future applicant state relies on non-judicial fact-finding material to establish genocidal intent.
States increasingly initiate genocide proceedings on the strength of United Nations human rights mechanisms rather than prior criminal convictions, a pattern also visible in South Africa v. Israel, presently pending before the same Court. How the Court treats the FFM’s material will therefore shape how future applicant states construct their evidentiary strategies, and how much weight investigative human rights bodies can expect their findings to carry once litigation begins. The question sits at an early and largely unsettled stage of the scholarship, commentators including Adnan (2026) have noted the institutional tension between the FFM’s investigative mandate and the Court’s adjudicative function. Moreover, the underlying question of exactly how demanding genocidal intent evidentiary’s threshold should be remains contested by both scholars and judges, as shown in Part II.
Neither the Genocide Convention nor the Court’s prior genocide jurisprudence addresses this question directly: Bosnia and Croatia arose against a background of prior criminal tribunal findings, not investigative material the Court had to assess for itself. This article identifies that gap and argues that it is resolved not by treating the FFM’s threshold and the Court’s standard as interchangeable, but by applying the Court’s own established practice of independently assessing the reliability of United Nations reports: the FFM’s “reasonable grounds” finding does not itself satisfy the Court’s standard, but the underlying material the Mission gathered may be independently weighed against it. The analysis adopts a doctrinal and case-based methodology, drawing on the relevant treaty provision, the jurisprudence of the International Criminal Tribunal for Rwanda (ICTR), the International Criminal Tribunal for the former Yugoslavia (ICTY) and the ICJ, the FFM’s own published methodology, and the oral pleadings delivered at the January 2026 merits hearings.
The remainder of the article proceeds in three parts. Part II traces the evolution of the “only reasonable inference” standard from its origins in ad hoc criminal tribunal jurisprudence to its adoption by the ICJ, and considers the debate about whether that standard is itself correctly calibrated. Part III examines the FFM’s own evidentiary methodology and the “reasonable grounds” threshold it applies to its findings. Part IV analyses the January 2026 oral pleadings, in which this precise tension was argued before the Court, and evaluates whether separating the FFM’s conclusion from its underlying material offers a doctrinally sound route through it.
II. The Evolution of the Intent Standard
Article II of the Genocide Convention defines genocide as any of a defined set of acts committed “with intent to destroy, in whole or in part, a national, ethnical, racial or religious group, as such.” (Genocide Convention, 1948, Art II). This mental element, dolus specialis, distinguishes genocide from other atrocity crimes and is widely considered the hardest element of the crime to prove(Ambos, 2009; Behrens, 2015). Confronting this problem directly in the first genocide conviction ever delivered by an international tribunal, the ICTR’s Trial Chamber described intent as a mental factor that is “difficult, even impossible, to determine” (Akayesu, 1998, para. 523). Direct evidence of genocidal intent, such as an explicit order or confession, is rare in practice, which forced tribunals to develop an alternative method of proof.
The ICTR’s response was not to abandon the intent requirement, but to permit its proof by inference: a conclusion a reasonable observer draws from the surrounding facts, rather than one stated directly by the perpetrator. In Akayesu, the Trial Chamber held that genocidal intent may be inferred from a defined set of factors, including the scale of the atrocities committed, their general nature within a given region, and the deliberate and systematic targeting of victims on the basis of their group membership while other groups were excluded (Akayesu, 1998, para. 523). This method, inference from a pattern of conduct, became the dominant approach to proving intent in subsequent international jurisprudence, precisely because direct evidence so rarely exists.
The ICTY’s Appeals Chamber subsequently narrowed this approach considerably. Ruling on the 1995 Srebrenica massacre, the Chamber held that where intent is inferred from a pattern of conduct, that inference must be the only reasonable inference available on the evidence (Krstić, 2004, para. 41). This raised the evidentiary bar substantially. A court applying this test could no longer treat genocidal intent as merely the most plausible explanation among several; the inference had to be the only one a reasonable observer could draw from the material before the court.
The ICJ imported this standard when it first ruled on the merits of a genocide claim. In Bosnia and Herzegovina v. Serbia and Montenegro, the Court held that for a pattern of conduct to be accepted as evidence of intent, it would have to be such that it could “only point to the existence of such intent” (2007, para. 373). It reaffirmed this position in Croatia v. Serbia, holding that inferring intent from a pattern of conduct requires that this be the only inference that could reasonably be drawn from the acts in question (2015, para. 148). Both judgments, however, benefited from an evidentiary environment the Court does not share in the same way in Gambia v. Myanmar: the ICTY’s own prior criminal findings on Srebrenica gave the Court in 2007 a body of judicially tested fact to draw upon, a foundation entirely absent in the Rohingya case (Adnan, 2026).
The “only reasonable inference” standard is not, however, universally accepted as correctly calibrated in the first place. Milanovic (2025) argues that the standard is not in fact specific to genocide, but rather a general evidentiary rule governing inferences drawn from circumstantial evidence in any context, and that it has become conflated in practice, though not formally in the Court’s own reasoning, with a beyond-reasonable-doubt threshold more familiar from criminal proceedings. On this view, part of the standard’s apparent severity reflects how it has been interpreted and applied by commentators and litigants, rather than anything strictly required by the text of the Convention or the logic of inference itself. This critique does not alter the analysis that follows, since the Court’s own practice in Bosnia and Croatia remains the operative benchmark against which any evidentiary material must be measured. It does mean, however, that the evidentiary translation problem addressed in this article sits within a broader and still unsettled debate about how demanding the required mens rea should be.
This is not only an academic disagreement, dissenting judgements has also highlighted the problems with such a demanding inference standard. Dissenting in Bosnia, Vice-President Al-Khasawneh argued that the majority had applied too exacting a threshold and had erred in declining to draw adverse inferences from Serbia’s refusal to provide key military documents, a position which he summarised as one of the key reasons why the Court reached the wrong result on intent (Bosnia and Herzegovina v Serbia and Montenegro, 2007, Dissenting Opinion of Vice-President Al-Khasawneh, paras 35, 62). Judge Cançado Trindade went further in his dissent in Croatia, warning that an unduly demanding threshold risks reducing genocide to a nearly impossible crime to prove and the Convention itself to “an almost dead letter” (Croatia v Serbia, 2015, Dissenting Opinion of Judge Cançado Trindade, para 94). On the other hand, there are also voices defending the high inference standard. Ambos (2025) challenged Milanovic’s argument directly, suggesting that the standard’s stringency may reflect a deliberate judicial choice specific to the exceptional gravity of a genocide finding, rather than mere conflation with the criminal standard.
III. The Fact-Finding Mission’s Own Standard
The UN Human Rights Council established the Independent International Fact-Finding Mission on Myanmar in 2017 to investigate alleged violations against the Rohingya and other minorities in Rakhine, Kachin and Shan States, after Myanmar’s own government had refused to cooperate with earlier UN human rights inquiries, making an independent third-party investigation necessary (UN Human Rights Council, 2018a, para 3). Unlike a court or tribunal, the FFM held no judicial or prosecutorial mandate; its function was strictly investigative. Its own recommendations were addressed to establishing individual criminal responsibility, through referral to the International Criminal Court or an ad hoc tribunal, and the prosecution of named senior officials rather than to state responsibility. A separate body, the Independent Investigative Mechanism for Myanmar (IIMM), was subsequently established in 2018 specifically to collect and preserve evidence for future criminal proceedings, and its material was also invoked during the January 2026 hearings; the analysis below, consistent with this article’s research question, concentrates on the FFM.
The Mission’s 2018 report set out its own evidentiary threshold explicitly. A finding was reached on reasonable grounds where a sufficient and reliable body of primary information, consistent with other available information, would lead an ordinarily prudent and impartial observer to that same conclusion (UN Human Rights Council, 2018a, para 6). Applying this threshold, the Mission drew on an extensive evidentiary base: 875 in-depth interviews with victims and witnesses, satellite imagery, and authenticated documents, photographs and videos (UN Human Rights Council, 2018a para 7). On this basis, the Mission concluded that “there is no reasonable conclusion to draw, other than the inference of genocidal intent, from the State’s pattern of conduct” (UN Human Rights Council, 2019, para 225), while separately recommending in its 2018 report that senior military officials be investigated and prosecuted so that a competent court could determine liability (UN Human Rights Council, 2018a para 87). The evidentiary conclusion and the prosecutorial recommendation are distinct: the former is a factual finding about the pattern of conduct, the latter reflects the Mission’s own institutional limits.
The distance between this threshold and the Court’s own is less a matter of evidentiary rigour than of institutional purpose. The FFM’s methodology already required verified and corroborated information, cross-checked against secondary sources; its threshold was not a bare credibility check. The difference lies instead in what each standard is asking: the FFM’s threshold governs whether an investigative body may state that a pattern or incident occurred and warrants further action, while the Court’s “only reasonable inference” standard governs a narrower and more specific question, namely whether, on the totality of the evidence, genocidal intent is the only inference a reasonable observer could draw, filtering out inferences that are not themselves reasonably supported by the evidence rather than requiring every conceivable alternative to be disproved. This inferential rule operates within, rather than as a substitute for, the Court’s broader requirement that allegations of such gravity be established to its full conviction on the record as a whole (Bosnia and Herzegovina v. Serbia and Montenegro, 2007). Notably, the Court itself has already engaged with this distinction once, without resolving it. At the provisional measures stage in 2020, the Court treated the Mission’s findings as relevant to establishing a plausible case for the purposes of indicating provisional measures, without converting them into a final judicial determination of intent (Adnan, 2026). That earlier, lighter-touch reliance now returns to the Court in a considerably higher-stakes form at the merits stage.
IV. Reconciling the Two Standards
Judge Charlesworth asked The Gambia’s counsel to address the consequence of the FFM’s “reasonable grounds” standard, which the Mission itself had described as lower than that required in criminal proceedings, for how the Court should approach the FFM’s evidence (Sands, in ICJ, CR 2026/18, 2026). Responding on The Gambia’s behalf, Philippe Sands argued that the difference between the standards was, in practical terms, of no real consequence, since the Court would in any event conduct its own independent assessment of the facts the Mission had found, regardless of the threshold the FFM itself had applied in reaching them (Sands, in ICJ, CR 2026/18, 2026). Counsel for The Gambia, Paul Reichler, reinforced this position at length, arguing that the volume, corroboration and institutional care behind the FFM and IIMM reports meant they merited “great evidentiary weight” before the Court (Reichler, in ICJ, CR 2026/18, 2026).
Myanmar’s counsel resisted this framing directly. Lead counsel Christopher Staker argued that the FFM’s reports, and the witness statements the Mission collected, did not meet the standard of evidence the Court should apply, while Myanmar’s Agent, Ko Ko Hlaing, characterised the Mission’s findings as a “condemnation without trial” and challenged their objectivity (Ko Ko Hlaing, in ICJ, CR 2026/7, 2026). Myanmar’s counsel Stefan Talmon further pressed the Court on the applicable burden and standard of proof, invoking the Court’s established requirement that findings of such exceptional gravity be fully and convincingly established before the Court (Talmon, in ICJ, CR 2026/23, 2026).
Myanmar’s counsel had earlier argued that The Gambia relied on only a small portion of the available IIMM material: the Mechanism had shared 42 witness statements with the parties, of which The Gambia submitted only 12 in support of its case (Staker, in ICJ, CR 2026/7, 2026). Reichler responded that the 12 statements were representative of the wider set, a characterisation the Mechanism’s own Head had endorsed in a sworn affidavit, and noted that Myanmar’s own submission of a further 19 statements from the same pool had, if anything, corroborated rather than undermined the FFM’s findings (Reichler, in ICJ, CR 2026/18, 2026).
Framed as a straightforward transplant of one threshold directly onto another, Myanmar’s objection carries real force. A finding reached under an “ordinarily prudent person” standard cannot, without more, satisfy a test demanding that intent be the only reasonable conclusion available on the evidence; the two thresholds measure different things. Reichler’s submissions, however, suggest a more defensible route through the problem than a straightforward transplant. Nothing in the Court’s jurisprudence requires it to adopt the FFM’s ultimate legal conclusion as such. In Bosnia, the Court set out the criteria it applies when weighing reports of this kind: the care taken in their preparation, the comprehensiveness of their sources, and the independence of those responsible for preparing them (2007, para. 230). Applied to the FFM’s reports, this is a framework for independently assessing reliability, not a rule requiring wholesale adoption or rejection of the Mission’s own conclusion. On this reading, the Mission’s “reasonable grounds” language describes only the confidence threshold the FFM itself required before publishing its own conclusion. It does not, and logically cannot, cap what a differently constituted body, applying a different and stricter test to the same underlying record, is entitled to conclude from that same evidence.
The Court’s own earlier practice lends some support to this separation. Treating the Mission’s findings as relevant without adopting them outright at the provisional measures stage in 2020 already implied a distinction between the underlying material and the Mission’s own conclusion, even if the Court did not articulate that distinction explicitly at the time. Extending the same separation to the merits stage, by treating the underlying interviews, imagery and documentation as evidence to be weighed afresh while placing no independent evidentiary weight on the FFM’s own conclusory finding, would offer a coherent way of applying the “only reasonable inference” standard to material that was never gathered with that standard in mind. Whether the Court is prepared to draw this distinction explicitly, and how much weight it ultimately attaches to the underlying record once it does, remains to be seen.
V. Conclusion
No provision of the Genocide Convention, and no prior ICJ genocide judgment, directly addresses how the Court should treat evidence gathered under an investigative threshold lower than its own inferential standard: this is a real gap in the Court’s evidentiary jurisprudence, opened by the growing reliance of applicant states on United Nations fact-finding bodies rather than prior criminal tribunals. That gap does not leave the Court without an answer. Its own settled practice of independently assessing the reliability of United Nations reports, applied consistently, resolves it: the FFM’s “reasonable grounds” threshold and the Court’s “only reasonable inference” standard are simply two different tests, developed for two different institutional purposes, applied to a shared body of underlying material. The FFM’s own conclusion cannot substitute for the Court’s independent judgment, but nothing prevents the Court from independently weighing the interviews, documentation and imagery that conclusion was built on, against its own considerably stricter standard. Both parties in the January 2026 hearings engaged directly with this distinction, even if neither framed it in precisely these terms: The Gambia argued for the evidentiary weight of the underlying material, while Myanmar argued that neither the material nor the Mission’s conclusion met the Court’s standard. The Court’s forthcoming judgment is expected to show whether this separation between conclusion and underlying material is one the Court is prepared to draw explicitly, and how far it is willing to extend it. Given the growing reliance of applicant states on United Nations investigative bodies rather than prior criminal tribunals, most immediately visible in the pending South Africa v. Israel proceedings, the Court’s answer in this case is unlikely to remain confined to it.
References
Adnan, K. K. (2026, May 11). From fact-finding to fact-proving at the ICJ: The Gambia v Myanmar and the Court’s coming reckoning on evidence, interventions, and intent. Opinio Juris. http://opiniojuris.org/2026/05/11/from-fact-finding-to-fact-proving-at-the-icj-the-gambia-v-myanmar-and-the-courts-coming-reckoning-on-evidence-interventions-and-intent/
Ambos, K. (2009). What does “intent to destroy” in genocide mean? International Review of the Red Cross, 91(876), 833.
Ambos, K. (2025). Proving genocide: A follow-up to Marko Milanovic. EJIL: Talk!. https://www.ejiltalk.org/proving-genocide-a-follow-up-to-marko-milanovic/
Application of the Convention on the Prevention and Punishment of the Crime of Genocide (Bosnia and Herzegovina v. Serbia and Montenegro), Judgment, I.C.J. Reports 2007, p. 43 (26 February 2007). https://www.icj-cij.org/case/91
Application of the Convention on the Prevention and Punishment of the Crime of Genocide (Croatia v. Serbia), Judgment, I.C.J. Reports 2015, p. 3 (3 February 2015). https://www.icj-cij.org/case/118
Application of the Convention on the Prevention and Punishment of the Crime of Genocide (The Gambia v. Myanmar), Provisional Measures, Order of 23 January 2020, I.C.J. Reports 2020, p. 3. https://www.icj-cij.org/case/178
Application of the Convention on the Prevention and Punishment of the Crime of Genocide (The Gambia v. Myanmar), Oral Proceedings, Verbatim Record CR 2026/7 (16 January 2026).
Application of the Convention on the Prevention and Punishment of the Crime of Genocide (The Gambia v. Myanmar), Oral Proceedings, Verbatim Record CR 2026/18 (26 January 2026). https://icj-web.leman.un-icc.cloud/sites/default/files/case-related/178/178-20260126-ora-01-00-bicc.pdf
Application of the Convention on the Prevention and Punishment of the Crime of Genocide (The Gambia v. Myanmar), Oral Proceedings, Verbatim Record CR 2026/23 (29 January 2026). https://icj-web.leman.un-icc.cloud/sites/default/files/case-related/178/178-20260129-ora-02-00-bi.pdf
Behrens, P. (2015). Between abstract event and individualized crime: Genocidal intent in the case of Croatia. Leiden Journal of International Law, 28(4), 923.
Convention on the Prevention and Punishment of the Crime of Genocide, 9 December 1948, 78 U.N.T.S. 277 (entered into force 12 January 1951), art. II.
Milanovic, M. (2025, September 18). Proving genocide. EJIL: Talk!. https://www.ejiltalk.org/proving-genocide/
Ministry of the President’s Office, Republic of the Union of Myanmar. (2026, January 19). Myanmar begins first oral argument at ICJ. https://presoffministry.gov.mm/en/news/22746
Prosecutor v. Akayesu, Case No. ICTR-96-4-T, Judgment (International Criminal Tribunal for Rwanda, Trial Chamber I, 2 September 1998). https://www.un.org/en/preventgenocide/rwanda/pdf/AKAYESU%20-%20JUDGEMENT.pdf
Prosecutor v. Krstić, Case No. IT-98-33-A, Appeals Judgment (International Criminal Tribunal for the former Yugoslavia, Appeals Chamber, 19 April 2004).
United Nations Human Rights Council. (2018a). Report of the Independent International Fact-Finding Mission on Myanmar (A/HRC/39/64). https://www.ohchr.org/sites/default/files/Documents/HRBodies/HRCouncil/FFM-Myanmar/A_HRC_39_64.pdf
United Nations Human Rights Council. (2019). Report of the detailed findings of the Independent International Fact-Finding Mission on Myanmar (A/HRC/42/CRP.5).
United Nations. (n.d.). Definitions of genocide and related crimes. Office on Genocide Prevention and the Responsibility to Protect. https://www.un.org/en/genocide-prevention/definition
