Colloquium: Addressing Racial Bias in Law Enforcement: An Interdisciplinary Approach to Developing Ethical AI Systems for Objective Threat Assessment

Title: Addressing Racial Bias in Law Enforcement: An Interdisciplinary Approach to Developing Ethical AI Systems for Objective Threat Assessment

Abstract: This research addresses the persistent challenge of racial profiling in law enforcement through an innovative interdisciplinary framework that integrates technical approaches to bias mitigation in AI, sociological understanding of systemic discrimination, and regulatory guidelines for responsible AI deployment. Despite decades of reform efforts, significant racial disparities persist in policing practices, as evidenced by recent Bureau of Justice Statistics data showing Black individuals are over three times more likely to experience use of force during police encounters. The research examines how AI systems, whilst potentially offering more objective assessments of suspicious behaviour, risk amplifying existing biases if developed without proper safeguards. The project’s novel contribution lies in its three-pronged approach: (1) developing advanced technical methods for bias detection and mitigation in multi-modal AI systems, including innovative adversarial debiasing techniques; (2) incorporating sociological insights into the multi-stage process of bias in policing to inform system design; and (3) creating practical implementation guidelines aligned with emerging regulatory frameworks such as the EU’s Artificial Intelligence Act. A sub-research component explores biometric applications using NoIR camera technology to measure blood pressure from the common carotid artery for airport security, demonstrating the practical application of ethical AI in high-security environments. This work responds to urgent calls for more holistic approaches to addressing algorithmic bias in high-stakes domains and aims to create law enforcement technologies that reduce rather than reinforce racial disparities.

Bio: Ahmed Eissa is a software engineer and academic researcher with extensive experience in AI integration and backend development. With an MSc in Software Engineering and a BSc (Hons) in Computer Science, he combines technical expertise in AI/ML tools, including TensorFlow, PyTorch, and OpenAI, with practical experience in architecting AI-powered systems. His professional background spans roles as Technical Team Lead at Brent General Construction Ltd and Digital Signage Software Engineer at W&J Linney Ltd, where he developed platforms used across the UK and EU. Eissa’s academic contributions include serving as an Associate Lecturer in Computer Science at Middlesex University, where he supervised research projects in AI, NLP, and machine learning, and co-authored peer-reviewed papers including “Crime Pattern Recognition Based on High-Performance Computing” and “Applying NLP for Crime Intelligence Analysis.” His interdisciplinary expertise in both technical implementation and academic research uniquely positions him to address complex challenges at the intersection of artificial intelligence, ethics, and practical applications in security contexts.