Poster #51 - Artificial Intelligence for Accountability: Reframing AI as a Tool for Civilian Oversight, Not Surveillance

• In Event: Poster Session I (posters 1 - 79)

Abstract

Artificial intelligence in criminal justice is typically associated with surveillance—predictive policing, facial recognition, and risk assessment—raising persistent concerns about bias and civil liberties. This poster reframes AI as a tool for accountability, demonstrating its potential to strengthen civilian oversight of policing. Using natural language processing (NLP), the project analyzes civilian complaint narratives to identify patterns of misconduct, including officer-specific trends, geographic clustering, and procedural justice concerns such as respect, neutrality, voice, and fairness. Anchored in the context of the Atlanta Citizen Review Board (ACRB), the study illustrates how AI-assisted analysis can enhance transparency, support data-driven investigations, and improve oversight outcomes. The framework prioritizes ethical safeguards, including bias audits, explainability, and human-in-the-loop review, ensuring AI augments—rather than replaces—human judgment. By repositioning AI as an instrument of oversight, this research offers a scalable, equity-centered model for integrating data science into police accountability systems.

Author

• Elycia Daniel, Clark Atlanta University