Abstract
International criminal prosecutions face growing evidentiary challenges due to the vast and heterogeneous digital material generated during conflicts, including user-generated videos, social media, and intercepted communications. This article examines the potential role of artificial intelligence (AI) in supporting investigations and trials before the International Criminal Court (ICC). Rather than replacing judicial decision-making, AI can assist with specific tasks, particularly in establishing contextual elements and linkage evidence central to international crimes. The article explores three main areas: AI-driven biometric tools for facial and speech recognition; AI techniques for detecting patterns of violence and mapping command structures; and applications during trial proceedings, including translation, summarization of testimonies, and cross-referencing evidence. While these technologies promise greater efficiency, they also raise critical concerns regarding bias, accuracy, explainability, and risks of automation bias. Errors such as false positives in facial recognition or misinterpretation of coded language in speech analysis may undermine prosecutorial duties under Article 54 of the Rome Statute. Moreover, if such tools are available exclusively to the Prosecution, disparities in access could compromise the principle of equality of arms when defense teams lack comparable resources. The article highlights safeguards such as disclosure protocols, independent audits, explainable AI, and training of court personnel to mitigate risks. It further considers implications for the admissibility of AI-generated evidence and the long-term challenges posed by the prosecution of international crimes resulting from the use of AI-enabled battlefield technologies.
html
