This presentation covers key AI applications across the criminal process, including predictive policing, risk assessment tools, facial recognition, AI-assisted forensic and digital evidence analysis, and automated case management. While these technologies promise efficiency and more consistent decision-making, they raise serious concerns for fundamental principles of criminal law and procedure, especially legality, fairness, proportionality, non-discrimination, transparency, and individualised justice. The lecture critically examines how probabilistic, data-driven outputs can influence legal reasoning and decision-making, sometimes producing “automation bias” and over-reliance on AI systems. Particular attention is given to evidentiary and procedural challenges, such as the admissibility and reliability of algorithmic outputs, the black-box problem and its impact on adversarial testing, and maintaining data integrity and chain of custody when AI tools filter or transform digital material. Human-rights implications are also addressed, including the presumption of innocence, equality of arms, privacy and data protection, and the risk of discriminatory outcomes driven by biased datasets. The lecture concludes by arguing for a rights-based, human-centered approach in which AI remains a decision-support tool under meaningful human oversight, supported by safeguards such as transparency, auditability, explainability, and stronger judicial and prosecutorial literacy.
Olga Kosevaliska (Fri,) studied this question.