Stanford Center for Racial Justice and Stanford SPARQ Summit on AI, Body-worn Cameras, and the Future of Policing I Andrew Broadhead

When Miami-Dade County police released the body camera footage of their stop and detention of Miami Dolphins star wide receiver Tyreek Hill,  America was watching. Everyone from policing experts to football fans pored through the videos trying to determine what went right and what went wrong. 

You might assume that some scrutiny of these incidents is standard practice—that police departments regularly review body camera videos to analyze interactions and improve practices. The reality is very different: most body camera footage sits in storage, unwatched and unanalyzed.

We believe this untapped data presents a crucial opportunity for research that could improve policing in America. This fall, in partnership with Stanford SPARQ, we convened the Summit on AI, Body-worn Cameras, and the Future of Policing, bringing together policymakers, law enforcement leaders, technologists, and policy experts to explore how we can harness this data to build fairer, safer, more equitable policing practices.

Over the past decade, body-worn cameras have become a standard tool in American policing. Adoption rates have soared from 32% of U.S. police departments in 2013 to 62% in 2020, with every department serving populations over one million now utilizing the technology. This technology represents the “largest new investment in policing in a generation,” with governments spending hundreds of millions of taxpayer dollars on implementation. 

The result is a vast collection of data documenting police practices and community interactions. Yet while these cameras capture countless hours of footage showing how police interact with the public—particularly in communities where Black and Brown Americans face disproportionately frequent encounters—most of this valuable data remains unanalyzed. Now, rapid advances in AI offer new possibilities for research leveraging this footage in ways that were previously impossible.

At this one-day summit at the Stanford Graduate School of Business’s Knight Management Center, more than 40 participants brought diverse perspectives on technology, research, and policing. San Francisco Police Chief Bill Scott and California’s Chief Deputy Attorney General Venus Johnson highlighted the potential of using AI to analyze body camera footage and improve law enforcement training. At the same time, civil liberties and privacy experts engaged in important discussions about the governance challenges this technology presents, emphasizing the need to carefully address privacy concerns as these powerful analytical tools develop.

The summit builds on the Center for Racial Justice’s work on Justice and Safety, particularly our extensive research on use of force and development of model policies. And now we’re taking this work further. We have partnered with Stanford SPARQ and an interdisciplinary team of researchers whose groundbreaking work analyzing police interactions has revealed racial disparities in how officers speak to drivers and their level of respect, how traffic stops escalate, and how training can shape these interactions.

Together, supported by a recent Hoffman-Yee research grant from Stanford’s Institute for Human-Centered Artificial Intelligence, we are developing research and policy frameworks to better realize the potential of body-worn cameras. This initiative will combine AI analysis of routine traffic stop footage from Bay Area law enforcement agencies with thoughtful policy development, aiming to create approaches that can benefit communities in California and beyond.

Dan Sutton is the Director of Justice and Safety at the Stanford Center for Racial Justice.