The coronavirus pandemic has brought out the weirdest and worst in surveillance technology. People are narcing on each other in new and creative ways, including flying drones to yell at people standing too close to each other, and neighbors across the U.S. calling the cops on each other for not keeping six feet apart. In the UK, cops set up an online system for snitching.
Artificial intelligence startups, of course, are getting in on the coronavirus-surveillance surge, with maps and tracking tools that source data about who's sick, who's social distancing , and where crowds are.
Voxel51, a startup led by a professor at the University of Michigan College of Engineering, is turning data from public closed-circuit television live cameras into a way to track whether people in some of the hardest-hit cities are obeying social distancing orders. The cameras currently in use are from several locations around the world, including New York's Times Square, Abbey Road in London, Fremont Street in Las Vegas, Seaside Heights in New Jersey, oceanfront Ft. Lauderdale, and two intersections in Dublin and Prague.
Using its own computer vision model, the system assigns a Public Distancing Index (PDI), an average measure of density in an area, every 15 minutes. It tracks vehicle, cyclist, and pedestrian traffic density. You can also adjust a chart graph to show the rate of cases and deaths in each region, relative to the PDI.
"This measure is non-invasive and does not use other data sources like mobile phone signals, which allow only approximate location estimates; instead we focus on specific centralized locations of interest in each city and literally watch what is happening," the Voxel51 website states.
Jason Corso, the founder and CEO of Voxel51, told Motherboard that the project is serious about protecting individual privacy. The data is "an aggregate statistical measure with no identifying information at all," he said. It's running on a major cloud vendor that is ISO/IEC 27001 certified [an internationally recognized standard of requirements for information security management systems], and integrates "state of the art redaction capabilities for privacy preservation," according to Corso. If the data were to be shared with law enforcement or health professionals, he said, it wouldn't contain identifying information.
"It helps people across the globe understand that they are not alone; it measures the specific human-mobility related motion in major public locations across the globe, with new feeds to be added regularly," Corso said. He hopes it has the potential to become a tool akin to the Hopkins Covid-19 case database, which tracks daily new confirmed cases and mortality rates by country. "For example, can computational social scientists develop a predictive measure that social activity has on case-rates in urban centers not yet heavily impacted or for future years?"
Streaming pre-existing CCTV footage within a pandemic context seems harmless enough—and might even be helpful for finding out when it's safe to visit certain areas, when we're allowed to go outside again. For those of us stuck inside severely-affected areas, there's a macabre voyeurism about watching people slowly mill around a desolate Times Square at several yards apart, or a lone Florida beachgoer wander through the frame. If the Voxel51 system is as secure as it claims to be, it could be a useful tool for health agencies to understand how distancing in typically-dense areas correlates with cases and severity.
But tools that use surveillance in an emergency, privacy experts warn, are dangerously prone to becoming the norm once the crisis is over. Employing CCTV data is certainly more reliable and respectful of individual privacy than, for example, cell phone location data or state-issued self-reporting apps. But even being watched by CCTV changes the way we operate in the world. When it's affixed to something as traumatizing as the pandemic we're living through, it could change how we move through the world when this is over, in ways we don't fully understand yet.