Digital Collection Tool: Collecting Witness Evidence Using a Culturally Sensitive Digital Investigation Tool

• In Event: Digital Criminology: Surveillance, Power and Accountability

Abstract

The advent of artificial intelligence (AI) and digitalisation has had a profound impact on the perpetration and documentation of large-scale atrocity crimes. The evolution of smartphones and other digital devices has ushered in a paradigm shift by increasing the availability of digital evidence, including photographs, videos, GPS coordinates, satellite imagery, and footage from drones, CCTV cameras, and other sources. This has facilitated the investigation and documentation of atrocity crime and has empowered survivors and witnesses to proactively engage in the pursuit of justice. However, concerns have been raised about the involvement of untrained personnel in the investigation of large-scale atrocity crimes. This paper proposes an AI-enhanced digital investigation tool that employs context-aware machine learning algorithms to guide investigators in the collection of intelligence and information. In comparison to other digital tools that already exist, this system integrates cultural knowledge bases through natural language processing and adaptive interfaces, while using federated learning approaches to maintain local data sovereignty. Further, it incorporates insights from legal psychological science to ensure that eyewitness evidence is collected in a way that facilitates memory retrieval and prevents suggestive influences on memory. This enables a more culturally sensitive, evidence-based and transparent method of information gathering during the digital collection of witness evidence, while ensuring investigative rigor and evidentiary standards.

Authors

• Michael Riegler, Simula Research Laboratory

• Nandor Knust, UIT Arctic University of Norway

• Annelies Vredeveldt, Vrije Universiteit Amsterdam