Utilizing Natural Language Processing to Identify Police Corruption in the United States of America (PR1 -Practice)

Deception detection has emerged as a vital interdisciplinary research area with far-reaching implications in law enforcement, judicial analysis, online fraud detection, and digital communication. While humans exhibit limited accuracy in identifying deceptive language, advances in Natural Language Processing (NLP) have enabled scalable and systematic analysis of linguistic cues linked to deception. Over the past decades, research in Automatic Deception Detection (ADD) has evolved from manually engineered linguistic features , traditional machine learning models to context-aware large language models (LLMs). This project aims to develop a LLM-based deception detection framework tailored for legal and police reports, addressing the challenges of data scarcity, ethical reliability, and reasoning transparency in high-stakes decision-making environments.