Polis launches TrustStat, the first multimodal AI system for analysis of police body-worn camera video
DALLAS — Polis announced the commercial release yesterday of TrustStat™, the world’s first multimodal AI system for the analysis of police body-worn camera (BWC) video. TrustStat integrates three kinds of advanced AI technology to enable police agencies and communities to examine key patterns in interactions between officers and the public. Identifying these patterns empowers agencies and communities with valuable, actionable data on police-community trust and other critical public safety issues.
The purpose of TrustStat is to help humanize policing by using AI technology to leverage the world’s largest source of data on police-community interactions: BWC video. Polis is initially deploying TrustStat in Dallas (TX), St. Petersburg (FL), and Kinston (NC), and is in discussion with other agencies and organizations nationwide.
Although most U.S. police agencies use BWCs, they have virtually no capacity to analyze the vast amount of data that the cameras record. As a result, the value of BWCs remains largely unrealized. TrustStat provides a transformative solution to this problem. Unlike less advanced technology that only analyzes transcripts from officers’ audio, TrustStat looks at the entire interaction using a combination of computer vision, natural language processing, and speech processing that analyzes behavior, semantics (what people are saying), and paralinguistics (how they are saying it). TrustStat is designed by social scientists, engineers, and police professionals to identify universal patterns of human interaction that are attuned to the rich diversity of communities and public safety agencies.
According to Polis CEO Jonathan Wender, “TrustStat’s AI technology follows the same patterns of observation and analysis that people use when they interact with each other.”
“To get a fair, complete, and accurate understanding of a situation,” says Wender, “you need to understand how people are acting, what they are saying, and how they are saying it.”
Following years of foundational research and development work funded by DARPA, NIJ, and BJA, Polis is expanding and commercializing TrustStat in collaboration with Microsoft. TrustStat is built on the Azure Government Platform, and is informed by Polis’ and Microsoft’s shared visions for the responsible application of AI to help foster greater trust, fairness, and equity in criminal justice services.
TrustStat is adaptively designed to meet a wide range of organizational and community goals ranging from optimized patrol operations and community engagement to more efficient, data-driven supervision and training.