Transforming Training through Body Worn Camera Analysis Using Multimodal Artificial Intelligence

NPI will assist the Spokane County Sheriff’s Office (SCSO), in partnership with Polis Solutions, in reviewing and evaluating the integration of body-worn camera (BWC) footage into training. Using advanced AI technology, the initiative aims to enhance deputies’ performance in de-escalation, community connections, and overall field interactions through data-driven training and evaluation.

Project Overview

Through collaboration with Polis Solutions and the National Policing Institute, SCSO will integrate BWC data into a modernized training model that spans deputies’ field training to in-service training. For three years, SCSO will use Polis Solutions’ “TrustStat,” a multimodal AI technology that analyzes both video and audio elements of BWC footage to identify behaviors and language related to de-escalation, use of force, and community interactions.

This project will systematically evaluate the effects of training at different career stages using BWC analysis. The initiative will develop training benchmarks, align them with job-critical skills, and implement real-time coaching protocols based on actual field footage and AI analysis. The outcomes will not only inform SCSO’s internal practices but also generate actionable insights and tangible strategies for law enforcement agencies nationwide.

Methodology

The project unfolds in three phases.

Year 1, NPI will define and align key behavioral features of effective policing, such as de-escalation and community interactions, with existing training curricula. A representative sample of BWC footage will be collected and rated both by SCSO instructors and by TrustStat separately to establish performance benchmarks.
In Year 2, these benchmarks will be used to improve training. BWC footage will be collected from deputies across training phases (field training and in-service), analyzed using TrustStat, and used to guide real-time coaching and training revisions. NPI will document the implementation and conduct surveys, focus groups, and observations to assess the impact.
Year 3, NPI will establish simple protocols for supervisors and instructors to review BWC footage that TrustStat has identified as significant. A working group will synthesize lessons learned and guide the project’s long-term sustainability and national scalability. Throughout the project, performance and outcome data will be collected and analyzed to assess changes in training quality, use of force, and community interactions.

Results

NPI and SCSO will deepen their understanding of the behavioral and communicative factors that lead to safe, respectful, and effective community interactions. By analyzing thousands of hours of BWC footage, SCSO hopes to identify specific patterns of deputy behavior and language that correlate with successful outcomes, including peaceful resolutions and high community satisfaction, in order to inform training strategies.

The project also seeks to evaluate how well training translates into field performance. By comparing BWC footage before and after training interventions, SCSO and its partners will learn which instructional approaches are most effective in real-world scenarios. Ultimately, the goal is to create a replicable model for training that is grounded in measurable field behavior and informed by cutting-edge technology—one that improves performance and strengthens community trust.