Assessing AI-Driven Predictive Policing and Its Impact on Community Trust

Thesis Abstract

The proliferation of artificial intelligence (AI) technologies in law enforcement practices, particularly predictive policing algorithms, has generated significant debate concerning their efficacy, ethical implications, and impact on community trust. This study investigates the extent to which AI-driven predictive policing influences community perceptions of fairness, legitimacy, and trust in law enforcement institutions. The primary aim is to assess the relationship between the deployment of predictive algorithms and changes in community trust levels, providing insights into whether such technological interventions foster improved police-community relations or exacerbate existing tensions. Specific objectives include examining community attitudes toward predictive policing, identifying factors influencing trust in AI-enabled policing, evaluating the perceived accuracy and bias of predictive algorithms, and exploring demographic variations in community responses. Employing a mixed-methods research design, the study integrates quantitative surveys and qualitative interviews to achieve a comprehensive understanding of community perspectives. The quantitative component targets a population of approximately 1,000 residents across urban neighborhoods that have implemented predictive policing programs in the past two years. Stratified random sampling ensures representation across age, gender, ethnicity, and socio-economic status. Data collection is facilitated through structured questionnaires measuring trust levels, perceptions of algorithmic bias, and explanations for attitudes toward predictive policing, with reliability tested using Cronbach’s alpha. Qualitative data are gathered via semi-structured interviews with 30 community members, law enforcement officers, and policymakers, aiming to contextualize quantitative findings and explore nuanced perceptions. Data analysis entails descriptive statistics to summarize community attitudes, followed by inferential techniques such as multiple regression analysis to examine relationships between predictive policing exposure and trust levels. Thematic analysis of interview transcripts identifies recurrent themes, perceptions of fairness, and trust-related concerns. The study is grounded in the procedural justice theory, which emphasizes fairness and transparency in law enforcement processes, and the Technological Acceptance Model (TAM), addressing acceptance and perceived usefulness of AI tools. Expected findings suggest that communities exposed to predictive policing are likely to report mixed perceptions while some acknowledge potential benefits in crime reduction, others express concerns about algorithmic bias, lack of transparency, and erosion of procedural fairness, which negatively affect trust. Demographic analysis is anticipated to reveal disparities based on ethnicity, socio-economic status, and prior experiences with law enforcement, influencing trust levels differently across groups. Furthermore, perceptions of algorithmic bias are hypothesized to mediate the relationship between AI deployment and community trust. This research contributes to the expanding body of knowledge on technological innovations in policing by providing empirical evidence on their social implications, particularly in fostering or undermining community trust. It offers actionable insights for law enforcement agencies seeking to implement AI-based tools responsibly, emphasizing transparency, fairness, and community engagement. The findings underscore the importance of aligning predictive policing strategies with community expectations and procedural justice principles. The study concludes that while AI-driven predictive policing has the potential to enhance crime prevention efforts, careful consideration of ethical and social dimensions is crucial. Recommendations include adopting inclusive policymaking that involves community stakeholders, increasing transparency around algorithmic logic, implementing bias mitigation measures, and conducting ongoing impact assessments. Future research should explore longitudinal effects of predictive policing on community trust and examine additional contextual variables influencing perceptions across different jurisdictions. Overall, this thesis emphasizes that technological advancements in policing must be accompanied by robust community-centered frameworks to sustain legitimacy and public confidence.