User-Generated Evidence: A Helping Hand for the ICC Investigation into the Situation in Ukraine?

[Konstantina is a researcher in the programme ‘Human Rights and International Criminal Law’ of the Ludwig Boltzmann Institute of Fundamental and Human Rights in Vienna, and PhD candidate at the Law Faculty of the University of Vienna focusing on the use of user-generated evidence in international criminal proceedings.]

Introduction 

Following Russia’s invasion in Ukraine in the last week of February, the international community has focused much of its attention in sanctioning the Russian Federation in an attempt to stop the war and minimise losses of human lives. After a failed attempt to adopt a draft United Nations Security Council (UNSC) resolution condemning Russia’s aggression against Ukraine, owing to former’s exercise of the veto power as a UNSC permanent member, States used the Uniting For Peace Resolution, which allowed them to take the matter up to the General Assembly. A resolution adopted on 2 March 2022 with 141 votes in its favour condemned Russia’s attack to Ukraine and expressed grave concern at reports of attack on civilian facilities, including residences, schools and hospitals as well as of civilian casualties, calling upon all parties to respect the relevant international humanitarian and human rights law provisions. Nevertheless, despite its political weight, the resolution is of non-binding nature. 

In parallel to these actions, a significant moment was marked in the history of international criminal law on 2 March 2022. The ICC Chief Prosecutor, Karim AA Khan QC, announced the opening of an investigation into the situation in Ukraine following receipt of referrals from 39 States Parties to the Rome Statute. The referrals enable the Prosecutor to proceed with an investigation into the situation without prior authorisation by the Pre-Trial Chamber, and to commence with the collection of evidence. In fact, as mentioned by the ICC Chief Prosecutor, he has already dispatched a team to the region to start gathering and preserving evidence of the commission of crimes falling within the Court’s jurisdiction. As user-generated content – content recorded by users on their personal devices, such as smartphones – depicting the commission of crimes in increasing, with researchers already delving into the work of gathering and studying potential evidence, it can be anticipated that the Office of the Prosecutor (OTP) will engage in the collection of this type of evidence as part of its recently announced investigation. Considering the ongoing discussions on the use of user-generated digital evidence in international criminal proceedings, some aspects deserve exploration. 

User-generated Evidence and Mis/disinformation

Prior to as well as during the ongoing Russian invasion in Ukraine reports have surfaced from investigative organisations, researchers and journalists pointing out to mis/disinformation spread on social media and on Russia state media outlets. In an attempt to curve the spread of disinformation, the European Commission banned Russian media outlets RT, Sputnik and their subsidiaries from the EU. Similar bans have been announced by Google’s YouTube, Facebook and TikTok. In addition, organisations such as Bellingcat have already started the process of debunking information, which can serve as a guidance for those reporting on the situation. 

The spread of mis/disinformation, including on social media, will likely present a challenge to the OTP and the task in this case will not be easy. The OTP first got familiar with user-generated evidence only five years ago in the Al-Werfalli arrest warrant. It has since relied on user-generated evidence in ongoing cases such as the one on the situation in the Central African Republic and in Myanmar/Bangladesh. The Ukrainian case will likely be, however, the first one involving such an overwhelming amount of user-generated evidence and will present one more test for the reliance on this type of evidence in proceedings before the Court. 

The main issue that might come up is the verification and authentication of user-generated content. The Berkeley Protocol on Digital Open Source Investigations, published in 2020, can assist in this regard. The Protocol combines best practices and creates comprehensive standards on open source investigations, including the perplexing procedures of evidence collection, verification, and analysis, which can serve as guidance in navigating this work and support the OTP’s accountability efforts. 

One important principle within the Protocol is data minimisation, i.e. avoiding the over-collection of evidence. This is particularly important in the case of Ukraine considering the speed of content creation and its increasing volume. As seen in the Syrian case, for instance, an approximate of 40 years’ worth of footage documenting committed crimes has been collected by Syrian Archive. Data minimisation can substantially reduce the challenges faced by the OTP in collecting and verifying a potentially similar overwhelming amount of user-generated content. However, while the Protocol provides a helpful starting point, it is dependent on organisations and researchers on the ground to adapt their work and conform to its guidance. 

Even in the scenario that organisations adhere to the standards of the Protocol, the ongoing Yekatom and Ngaïssona case before the ICC signposts to further issues that might arise during trial proceedings when user-generated evidence is relied on. In this case, the defence and its witnesses have challenged the evidence as rumours and fake news. Considering that Russian media already qualifies information and reports about the events in Ukraine as false and misleading, comparable defences might be raised in future proceedings and it remains to be seen how the ICC will respond and the weight it may attach to user-generated evidence. 

User-generated Evidence and Biases’ Mitigation

Another possible challenge is the mitigation of biases. As user-generated evidence is exclusively dependent on the availability of technologies, it could result in biases in evidence collection. For instance, cities are facing blackouts and disruptions in internet connectivity, which might affect the documentation of crimes in some areas with user-generated content not reflecting the actual broader situation. This will require further caution during the evidence collection and analysis to ensure, on the one hand, the utility of user-generated evidence in proceedings and, on the other hand, the representativeness of types of crimes and victimised populations.

Moreover, as actors involved in evidence collection usually aim at promoting accountability, it is likely that more inculpatory rather than exculpatory evidence will be gathered. This could, in turn, affect the equality of arms as defence teams might not have the same opportunities to gather exonerating evidence to support their claims and to counter evidence presented by the OTP. As the ICC is obligated to ‘investigate incriminating and exonerating circumstances equally’, additional efforts shall be dedicated in identifying user-generated evidence that could serve the rights of the defence. 

Owing to the increasing volume of user-generated content, machine learning and artificial intelligence (AI) will probably be used as tools for evidence identification and verification. While such technologies can be helpful, they are not perfect. The use of AI as means to identify evidence in the abundance of information available is subject to technical biases. Algorithms tend to prioritise information which receive bigger attention, posing difficulties to the discovery of some crimes and leading to the over-documentation of others. This could affect investigators’ perception of events leading to selection biases and possibly leaving populations further marginalised. Lastly, algorithmic biases might affect the analysis of recovered evidence. Since algorithms are informed by training data, they can have inherent biases leading to discrimination. These issues will need to be addressed by the OTP during its investigation to ensure the impartiality of evidence and the fairness of the proceedings.

Conclusion

The initiation of the OTP investigation into the situation in Ukraine has raised hopes of accountability for the crimes being committed in the ongoing conflict. Owing to the sheer amount of user-generated content documenting the perpetration of crimes, the OTP will possibly resort to the collection of and reliance on user-generated digital evidence. Even though this type of evidence can be a helping hand for the Prosecutor, it requires careful use in order to safeguard the integrity and fairness of proceedings.