Artificial Intelligence Adoption in Forensic Science for City Security in the UAE

Abstract

The rapid adoption of Artificial Intelligence (AI) in forensic and security-related practices has transformed investigative processes, until now its effectiveness depends on organisational, ethical, and technological conditions. This study proposes and empirically tests an AI–TFPs (AI–Transformational Forensic Processes) framework to examine how AI integration, ethical and legal requirements, and technological advancements influence forensic outcomes through key mediating mechanisms: Adaptation and Learning, Iterative Evaluation, and Stakeholder Engagement. Using Partial Least Squares Structural Equation Modelling (PLS-SEM) and data collected from 403 professionals in the United Arab Emirates, the study evaluates both the measurement and structural models. The results confirm strong construct reliability, convergent validity, and discriminant validity. Structural model findings reveal that AI integration, ethical–legal requirements, and technological factors do not directly improve forensic outcomes. Instead, their effects are primarily transmitted through organisational adaptation and learning and stakeholder engagement. Adaptation and Learning emerge as the strongest predictor of forensic outcomes, followed by Stakeholder Engagement, while Iterative Evaluation shows no significant direct or mediating effect. The model demonstrates substantial explanatory power (R²), meaningful effect sizes (f²), strong predictive relevance (Q²), and acceptable model fit. Overall, the findings highlight that successful AI adoption in forensic contexts requires more than technological deployment where it depends on continuous learning, ethical governance, and active stakeholder involvement. The study provides practical and theoretical insights for policymakers and law enforcement agencies seeking to enhance forensic effectiveness through responsible AI implementation.