Immersive Technology to Help Advance Public Safety
What does the future of public safety look like? Civil rights groups, law enforcement agencies, researchers, and community representatives are all working to try to answer this question. Part of the solution will involve training public safety personnel, including law enforcement officers and other first responders, in a range of new communication and de-escalation skills. While early research on immersive, scenario based training is promising, doing this research is challenging and inaccessible to many of the institutions that can provide value. We are missing important voices. And once developed, scaling this training to the over 18,000 police departments in the U.S. is prohibitively expensive.
Recent advances in virtual reality (VR) have demonstrated the potential for technology to create scalable opportunities for law enforcement and other public safety professionals, particularly in the realm of education. In training contexts, VR can create a uniquely immersive experience, employing heightened tensions to build critical skills in an environment that mimics the same physiological responses as those generated in real-world interactions.
Through collaboration with a diverse group of civil society organizations, academics, researchers, and law enforcement in the U.S., we explored whether VR could advance de-escalation training by enhancing communication and critical thinking skills. The resulting work is Trainer — a VR platform that combines recent advances in voice recognition, natural language processing, and VR to provide law enforcement instructors, and criminal justice scholars, with an immersive, realistic environment to train and evaluate officer performance.
How Trainer works
Once a user puts on the VR headset, they’re immersed in a virtual world and experience interactive scenarios with responsive virtual characters of different races, ages and genders. Scenarios are modeled on everyday police interactions such as domestic disputes or motor vehicle stops to hone officers’ problem-solving skills. The characters respond to questions and commands in real time, and their answers change according to the user’s language and behavior, including body positioning and interaction with objects in the virtual environment. Instructors can observe and evaluate officer performance in scenarios with different contexts and stressors.
The fluid exchanges between users and virtual characters are a result of integrating with Dialogflow — a natural language understanding platform that supports building speech interfaces into apps and interactive voice-response systems. Dialogflow works as a classifier to take user speech as an input and categorize it based on likely user intent. It can be expanded to accommodate new language over time, including regionalisms and novel responses, to increase the naturalness and verisimilitude of the interactions. This enabled us to add new layers of depth to the virtual characters as we tested the system with officers around the country.
An informed approach
From the inception of this project, Trainer has relied on the research of and collaboration with a diverse group of civil society organizations, academics, researchers, civil rights activists, and law enforcement in the U.S. The input of these groups has been critical in addressing the question of whether technology can help make policing safer for all. We turned to a variety of organizations, including The Leadership Conference on Civil and Human Rights and the National Action Network, to understand how histories of police misconduct and past violations of civil rights influence law enforcement training today — a necessary step toward determining whether and how training can be improved through VR. Similarly, over the course of the project, we sought to validate our work through user testing and research with several police departments across the country including the Camden County Police Department, Jersey City Police Department, and Stockton Police Department.
Leaders within these organizations have expressed their enthusiasm for Trainer’s potential and the approach to advancing VR training.
Ebonie Riley, National Action Network’s Washington, DC Bureau Chief, said that she is “proud to have been included in a broad, diverse coalition of advocates, scholars, policy experts, and law enforcement who added input to Jigsaw’s ground-breaking technologies. Jigsaw’s ability to merge technological advancements, cultural dimensions, and implications for civil rights and social justice in a world of growing inequality is the definition of bringing all voices to the table.”
James Shea, Jersey City’s Director of Public Safety, said, “The immersive training takes de-escalation tactics to a new level, providing various situations, environments, and emotions, just as our police officers experience every day in every community. Nothing studied in a textbook or shown on a screen can compare to the promise of Jigsaw’s virtual reality training.”
Where we go from here
Understanding Trainer’s efficacy is an important next step in the journey to leveraging it for future use as an instructional tool. Beyond training, the technology offers a controlled environment for research. The platform can provide researchers with the opportunity to study officer behavior and explore how to reduce violence between police and communities. Trainer can be modified and expanded to support study goals by adding scenarios, changing characters’ appearance or behavior, and modifying environmental stressors to further analyze a wider variety of interactions. We believe this creates new opportunities to contribute to the broader goal of advancing public safety.
Today we are transferring Trainer’s technology to a new consortium of academic institutions to drive research on the platform. These institutions include University of Cincinnati, Morehouse College, University of Maryland, and Georgetown University. Each of these programs presents a unique set of research questions to the issue of equitable public safety and seeks to leverage Trainer in their ongoing research.
- The Center for Police Research and Policy at the University of Cincinnati will study how Trainer could improve outcomes between police and communities, building upon their first-of-its-kind study on the impact of police de-escalation training on officer behavior in the field.
- At Morehouse College, the Culturally Relevant Computing Lab and the National Training Institute on Race and Equity will study Trainer’s effect on officers’ empathy toward Black Americans and will leverage the technology to strengthen anti-bias trainings for law enforcement, in which students can serve as research assistants and training co-facilitators.
- The Lab for Applied Social Science Research at the University of Maryland will leverage Trainer to focus on the factors that determine negative outcomes between police and communities such as understanding what virtual scenarios can teach us about bias in policing.
- At Georgetown Law’s Center for Innovations in Community Safety, the Active Bystandership for Law Enforcement project will incorporate Trainer into its core work of teaching officers how to recognize harmful behavior by their fellow officers in the field — and intervene to prevent it.
We always believed that to be an effective technology for real-world use, Trainer would need to be in the hands of those who have the expertise in evaluation and research to effect change. With the help of these partners, our hope is that Trainer will help police officers improve critical skills and better inform academics, researchers, and practitioners on successful deployment of VR training, and ultimately drive progress towards fairer policing in communities most impacted by police violence.
By Kevin Rabsatt, Software Engineer, Maddy Hoffman, Product Manager, and Sameer Syed, Partnerships Manager at Jigsaw
