Digital Evidence Before the ICC: Lessons from the Yekatom/Ngaïssona Judgment

[Dr. Emma Irving, M.A., LL.M. and Sabrina Rewald, J.D., LL.M. are co-founders of the Fénix Foundation, a non-profit leveraging technology to advance peace, justice, and accountability, and consultants in international criminal law, human rights and technology]

The authors led the research and development of the Leiden Guidelines and the Hala Protocol.

On 24 July 2025, ICC Trial Chamber V found Alfred Yekatom and Patrice-Edouard Ngaïssona guilty of a list of war crimes and crimes against humanity committed in the Central African Republic between 2012 and 2014. Following the trend set by other recent ICC judgments, digital evidence plays a notable role in the judgment. This post contains preliminary thoughts on the judges’ approach to two types of digital evidence: Facebook Materials and Call Data Records (CDR). 

The prominence of Facebook Material and CDR in the judgment reflects their prominence in the parties’ cases: the Chamber recognised as formally submitted approximately 943 items related to Facebook evidence and approximately 104 items related to CDR evidence. Facebook exchanges (including ‘private messages…between users’ and ‘screenshots of posts visible on the Facebook user’s profiles’ [para. 144]) were used to–among other things–support Prosecution arguments regarding Mr. Ngaïssona’s presence and activities in certain areas (para. 464) and his coordinating role in the Anti-Balaka (para. 435). CDR (‘comprising call data, cell site items, and call sequence tables’) were used as evidence of the existence and timing of contact between particular individuals (e.g. fn 5579).

The Chamber addresses questions raised by the Facebook Material and CDR early in the judgment in a distinct section on ‘specific evidentiary considerations’ (p. 47 onwards). In some respects, the approach outlined by the Chamber is consistent with previous (digital) evidence case law; in others, it provides much needed guidance on questions left open by previous cases. 

Confirmation of Existing Approaches

Regarding consistency with previous case law, there are two points that jump out. First, the importance of author testimony. The Chamber notes, in relation to Facebook exchanges, that such material will have little probative value if not accompanied by ‘a statement or testimony from an individual who participated in a Facebook exchange or authored a post, or otherwise has information about the contents’ (para. 154). Indeed, the Chamber notes that Facebook Material alone (which can include data provided by Facebook itself, including IP addresses) is not enough to establish the authorship of a Facebook post in the absence of testimony from the author or someone who participated in the exchange. This reasoning aligns with prior ICL case law on the importance of author testimony, for example from the ICC (see Al Hassan, para. 726, fn 2175), from the ICTY (see Delalic, para. 22), and from the ICTR (see Renzaho).

Second, the Chamber is consistent in reiterating the importance of corroborating evidence, despite there being no legal requirement to do so per RPE 63(4). The Chamber stressed the importance of assessing evidence holistically and identified a number of points that neither Facebook nor CDR materials alone could substantiate. For example, Facebook IP addresses are not enough to prove geographic location (para. 151), and CDR materials are not enough to attribute a phone number or handset to a given individual or to prove that a conversation between two identified individuals actually occurred (paras 188-189). In these instances, corroborating information will be needed. This approach accords with prior ICL case law regarding digital evidence, for example, concerning videos (see Al Hassan, para. 1031, fn 3458), intercepted communication (see Krstić, paras 114-116), and photographs (Ntaganda, para. 68). Applied specifically to CDR and Facebook Material, the judgment offers both confirmation and guidance anew.

Finally, the Chamber reinforced the strong indication of authenticity when data is provided by a relevant commercial actor. For example, corporate watermarks on CDR have been considered inherent indicia of authenticity (see Bemba et al, para. 219, fn 232). In the Yekatom/Ngaïssona  case, the relevant commercial actors were Facebook and four telecommunication companies, with the data obtained by the competent (Irish and CAR) authorities (paras 147-149; 161). The Chamber rejected arguments that the Facebook Material and the CDR had failed to be duly authenticated. It cited ‘no reason to doubt’ that the Facebook (para. 148) and CDR (para. 182) evidence corresponded with what the authorities had obtained from the commercial actors, nor to doubt ‘the validity of the respective chains of custody as reflected in the metadata’ of the CDR (para. 182). Pertaining to the Facebook Material, the Court deemed ‘mere submissions’ of alleged tampering to be speculative (para. 148).

New Guidance on Digital Evidence Issues

In addition to confirming existing points, the judges of Trial Chamber V did not shy away from tackling technical questions. As a result, there are some new insights that we can take away from the judgment:

First, even in the current uncertain information landscape, there is not a presumption that all Facebook messages are misinformation until proven otherwise. While recognizing the Facebook Material to have little probative value absent corroboration, the Chamber rejected the broad presumption that all Facebook Materials should be considered unreliable due to the potential for Facebook users to post misinformation or disinformation; instead, the Chamber stated that a case-by-case determination of the Material is appropriate (para. 154). In doing so, the Chamber took a stance against the digital-era harm of the liar’s dividend—the idea that the prevalence of false information (lies) makes people more skeptical of all information, including the truth, thereby helping perpetrators to evade accountability (the dividend).

Second, CDR and cell site items can ‘generally be presumed reliable’. While, similarly to the Facebook material, the Chamber found the CDR to have low probative value, it weighed the following features in favor of their reliability: that CDR information tends to be generated, transferred, and verified via automated processes ‘in the ordinary course of business’ (para. 170); that their collection is overseen by people with sufficient knowledge or expertise; that CDR in their raw form cannot be altered, which guarantees authenticity; and, that CDR accuracy serves the cell network’s best interest. This reliability presumption can however ‘be rebutted by […] showing specific examples of irregularities in the collection, storing, formatting and/or transmission of the evidence.’ (para. 183)

Similarly, the judges’ clarifying distinction between ‘reliability’ and ‘authenticity’ as they pertain to CDR evidence is a welcome reflection. The Chamber asserts that questions as to the nature, authorship, and process of creation of CDR evidence are a matter of reliability, not authenticity. It states ‘the potential disparities in formats of the CDR or the fact that some may have been generated for billing, tax or litigation purposes do not have an impact overall on the authenticity of the CDR. […T]hese differences do not necessarily mean that some CDR ‘are more genuine’ or more authentic than others.’ (para. 182).

An Updated Lens on the ICC’s Submission Approach to Evidence 

As a final reflection, the judgment is an encouraging indication of how the ICC’s ‘submission’ approach to evidence, lauded for its convenience, can incorporate the discursive advantages of the admission approach. The submission approach entails the formal submission of all items of evidence, deferring the assessment of admissibility to final deliberations, unless an earlier evaluation is procedurally required. In a previous post, the authors cautioned that the ICC’s transition to the submission approach appeared to result in a diminishment of discussions by Chambers on the treatment of digital evidence. Such discussions—that would generally be found in decisions on the admission of evidence—are crucial for providing guidance for practitioners and investigators, and help to inform the creation of resources such as the Berkeley Protocol, the Leiden Guidelines, and the Hala Protocol. In the Yekatom/Ngaïssona judgment, the Chamber addressed its adoption of the submission approach head-on, reflecting that ‘it assessed freely all evidence in accordance with Rule 63(2) of the Rules and considered all the evidentiary criteria for each item of evidence submitted, although it may not necessarily discuss in the judgment every submitted item’ (paras 130-133). The Chamber then proceeded with its extensive discussion of the evidentiary considerations pertaining to Facebook and CDR materials, demonstrating how the submission approach can balance the convenience of limiting admissibility decisions throughout the trial (and the delays this can cause), with the need for detailed discussion of evidentiary standards (and the insights this can deliver). In this respect, the authors hope the judgment can serve as a template for future cases. 

Conclusion

Overall, as well as being (largely) a victory for victims, the judgment better equips investigative and legal practitioners, both Defence and Prosecution, to handle what is now a standard and crucial category of evidence: digital evidence.