DIGITALIZE IT: DIGITAL EVIDENCE AT THE ICC
Tal Mimran & Lior Weinstein
The International Criminal Court (ICC, or the Court) first accepted digital evidence in a legal proceeding in 2013 during the prosecution of Al Faqi Al Mahdi for ordering the destruction of the Timbuktu shrines and mosques in Mali. Since that case, there has been growing discussion of the need for greater inclusion of digital evidence in international criminal law processes, notwithstanding lingering concerns.
The importance of modernizing the ICC’s evidentiary process is evident in the ongoing conflict between Russia and Ukraine. The Prosecutor of the ICC, Karim Kahn, acknowledged the need to modernize when he committed to ensuring relevant forensic and digital materials are collected in a manner that strengthens their admissibility in future proceedings before the ICC. The consideration of digital evidence received a further boost when the ICC, under the stewardship of Khan, launched an advanced digitalized evidence submission platform known as OTPLink.
At the same time as these developments, the Palestinian Authority (PA) Assistant Foreign Minister for United Nations and its Specialized Agencies, Omar Awadallah, indicated that the ICC intends to launch an online complaint platform for Palestinians that will enable them to submit pictures and videos of crimes under the jurisdiction of the Court.
This post discusses the ongoing and developing process of digitalization at the ICC, specifically in the context of submission and reliance on digital evidence. We first refer to developments and case law dealing with digital evidence, with a focus on the authenticity of evidence under the ICC’s existing framework. We then discuss more broadly the pros and cons of a digitalization process at the Court. Finally, we will suggest mitigating measures.
Digital Evidence and Authenticity at the ICC
In the context of the Russian invasion of Ukraine, many have highlighted the role private online documentation of alleged crimes could play in the promotion of accountability. The Court has already received numerous submissions that include digital evidence related to the Russia-Ukraine conflict. Recognizing the potential significance of this material, the Council of Europe held a webinar on the collection of digital evidence in November 2022.
Digital evidence can be extraordinarily probative as evidence of criminal intent and acts. Its compelling presentation of reality (through sound and video) can convincingly establish both actus reus and mens rea. At its best, digital evidence, such as material found on social media, can provide a unique glimpse into the intent of an accused. When used in conjunction with traditional forensic evidence and witness testimony, digital evidence can helpfully diversify the bases for factual claims brought before the Court.
At the same time, there are questions as to the best method to promote credibility and verify the authenticity of digital evidence. Even potentially strong digital evidence, such as a video clip, can be insufficient (because of what is or is not visually captured, or because it could be taken out of context). Intentional manipulation or falsification of alleged digital evidence, such as the creation of deep fake videos, can also be a problem, one that goes hand-in-hand with technological leaps.
Admission and Evaluation of Digital Evidence
The use of digital evidence is, nevertheless, not novel at the ICC. In fact, important guidance can be found in the Court’s jurisprudence in the recent decade, especially relating to the admission of evidence under the Rome Statute.
Article 69(4) of the Rome Statute encompasses a three-tiered test for the admission of evidence. First, the Court examines whether the evidence is prima facie relevant (see here para. 9). Second, the Court considers whether the evidence contains probative value as determined by factors including: reliability; trustworthiness; credibility; and authenticity (para. 26). Finally, the Court evaluates whether the factors at hand have a prejudicial effect on the accused’s fair trial rights.
In the case of digital evidence, the Court has noted several indications for authenticity, for example, “self-authentication” within the evidence, such as geolocation data (p. 113) and metadata, mainly using expert witnesses (Al-Werfalli arrest warrant, para. 18) and external validation or corroboration from additional documentation or testimony (Ntaganda trial judgment, paras. 59-60).
Yet, some claim the ICC is ill-adapted to assess digital evidence (see here and here). On a technological level, it has been suggested that the Court uses weak hash-algorithms and systems that do not allow the collection of metadata—such as location, altitude, and barometer readings—rendering the authentication process incomplete or insufficient. Legally speaking, questions also arise as to adaptability or to indications implemented with regard to “classic” evidentiary doctrines, like those used with conventional forensic evidence.
Potential Benefits of the Digital Platform
Concerns notwithstanding, digitalization can play an important role to support testimony in a legal proceeding. For example, a reliable satellite image, if credible and authentic, can strongly corroborate a witness’s claims or descriptions of a physical environment. Additionally, the use of technology releases the Prosecutor from relying solely on traditional forensic evidence and testimony, which are often difficult to gather in international criminal investigations. This can be due to geographical and temporal challenges, as well as lack of cooperation by a State whose officials are under investigation.
Questions relating to the use of digital evidence at the ICC became even more urgent in May 2023, when the Office of the Prosecutor (OTP) launched OTPLink. This online platform allows victims, witnesses, non-governmental organizations (NGOs), and others to digitally submit evidence relating to international crimes directly to the OTP. The platform is part of a larger digitalization effort at the Court and the OTP. As mentioned above, OTPLink was followed by Project Harmony, which aims to use artificial intelligence (AI) to analyze evidence and include rapid pattern identification, video and image analytics, facial identification, and more. If reports are correct, the online complaint platform for Palestinians is yet another step in this larger digitalization process.
Interestingly, both OTPLink and Project Harmony were developed and introduced with the support of the European Union and the assistance of companies from the private sector, including Microsoft. While such support is encouraging in terms of multi-stakeholder partnership, the Court should not be tempted to circumvent cooperation and consent of the States involved in the situation investigated. Otherwise, the digitalization process may drag the Court from its original vision and mission of complementing domestic criminal jurisdiction in a cooperative form with the Member States.
Another aim of the digitalization effort is to extend the Court’s outreach relating to situations. Such outreach is likely to focus on “increasing knowledge about the Court and providing correct information about its work, jurisdiction and mandate.” Broadly speaking, outreach promotes the wider goal of victim participation in proceedings at the ICC as part of a restorative justice rationale. Direct relationships with the civilian population reduce the role of intermediaries that might manipulate victims’ perspectives improperly or harm the restorative effect of the trial and victim participation.
The need for improved inclusion and participation by victims is not theoretical. In 2020, the Independent Expert Review of the ICC released a Final Report, which identified many procedural difficulties regarding the effective participation of victims. An accessible online platform to submit evidence can contribute to the role of victims in the procedure, increase its effect and allow presentation to the OTP of victim concerns and perspective on the situation and help navigate the focus and results of the investigation.
In sum, digital platforms of communication with the Court promise to simplify the submission of evidence and may open the door to a more inclusive adjudicative process. As such, digitalization has a positive, democratizing potential. For this positive goal to be reached, the platform should be made approachable in practical and technological terms.
Potential Drawbacks of the Digital Platform
There are some problems, naturally, with reliance on digitalization in the work of the ICC. One of the predominant risks is the use of misinformation in a legal proceeding. Indeed, edited or manipulated evidence can easily be submitted to the Court. And the Court may not be equipped or able to detect it in all cases. The Court could certainly be fed a false narrative by State-sponsored campaigns to influence its results.
This concern is more troubling when AI is used to analyze evidence provided to the OTP or the Court. In training stages, such incidents might generate future biases in the examination of other available evidence, with the Court turning its sights toward larger trends, or even failing to rule out evidence. Such problems may grow in the face of high volumes of evidence submitted via OTPlink (which allows users to upload up to 1000 files in each submission).
Similarly, digital evidence is open to algorithmic biases (e.g., in relation to racial, ethnic, or gender issues). For example, algorithm-based content moderation might lead to an uneven exclusion of content submitted to the Court. In the case of systems for facial recognition, the model might be trained on less diversified data sets and lead to inaccurate and biased recognition (as occurred in the context of domestic criminal enforcement against non-Caucasian figures). Gender biases might also occur, due to gaps in documentation of harm to men and women, or due to societal norms.
In a high-profile conflict, such as the Israeli-Palestinian conflict, which is both highly documented and full of disinformation, these concerns are of even greater relevance. In addition, there is also the concern over methodological and technological blind spots that should be considered when assessing evidence submitted by digital platforms.
The need to balance the opportunities and risks from growing reliance on digital evidence and digital outreach platforms invites thoughts on mitigation. We believe that although AI can be extremely helpful to investigations and prosecutions, prudence is due in the early stages of its use. In particular, AI should chiefly be used to assess external and internal indications of reliability such as metadata and to analyze indications of authenticity.
In addition, given algorithmic biases, as well as technological and methodological blind spots, corroboration or aggregation of digital evidence with forensics and testimony should be preferred. Naturally, the standard of proof of each piece of digital evidence should be carefully assessed. The weight given to digital evidence should be decided based on questions concerning chain of custody, problems with which invite either a conclusion of low probative value or of exclusion from consideration. As such, it is preferable to use digital evidence to strengthen or corroborate allegations based on more “traditional” evidence to minimize the prejudicial effect on the accused’s fair trial (art. 69(4), Rome Statute).
Further, technical training is important both in terms of understanding the mechanisms through which digital evidence is obtained, and in terms of potential biases and risks. Such training should be given to OTP staff and the judges of the Court to better mitigate the drawbacks of digital evidence.
Finally, safeguards should be considered with the aim to cope with blind spots. For example, technical or methodological safeguards should be implemented to better improve explanations of the algorithmic or automated decision-making process. Such safeguards can involve keeping a human in the decision-making loop, or technical oversight tools that limit the overuse or over-reliance on digital evidence based on algorithmic or automated decision-making.
In the Russia-Ukraine context, as well as the Israeli-Palestinian conflict, the ICC is faced with heated, and highly politicized situations. In these contexts, there are high volumes of testimony and evidence gathered by the warring actors and by third parties, including NGOs, international organizations, media and more. An overly dense and complex digital evidence pool can be a double-edged sword that distracts the Court from its goals, pollutes the criminal process, and infringes on fair trial rights.
In our view, the trend of digitalization in the ICC holds great potential. To fulfil this promise, there is a need to balance between the hopes and fears of this new means of supporting international criminal litigation. This can be done through prudent incorporation of AI in evidentiary analysis, dealing with misinformation and falsifications, and coping with algorithmic biases. Such steps, in turn, can improve judicial decision making and promote restorative justice, and further democratize criminal procedure.
Dr. Tal Mimran is an adjunct lecturer at the Hebrew University of Jerusalem and at the Zefat Academic College. He is the Academic Coordinator of the International Law Forum of the Hebrew University, and the Research Director at the Federmann Cyber Security Research Center in the Law Faculty of the Hebrew University.
Lior Weinstein is a Master‘s student of international law (LLM) at the Hebrew University, Jerusalem and a Researcher at the Tachilit Policy Center in the fields of Law and Technology and International Law.
ד"ר טל מימרן
הוא ראש תוכנית "אמנה חברתית לעידן הדיגיטלי" במכון תַּכְלִית, חוקר ומרצה בתחומי המשפט הבינלאומי והסייבר.