What did the Seattle Police Department do when an activist requested their entire archive of patrol car videos — all 1.6 million videos? For the hometown of Bill Gates and Jeff Bezos, the answer was easy: Seattle’s cops went to the computer nerds.
Law enforcement agencies are promising body cameras will bring a new era of accountability by capturing cops’ every interaction on film in Seattle, Los Angeles, New York and other cities encouraged by Obama’s promise of $263 million in funding. But all that video presents a technical problem: how can a department possibly sort and release so many hours of footage? Stepping up its commitment to transparency and collaboration, Seattle’s police asked 80 local tech wizards from Amazon, Microsoft and Evidence.com to streamline the disclosure process at its first department-sponsored hackathon earlier this month.
“We’re having a conversation about transparency and privacy. How do the two intersect?” Sgt. Sean Whitcomb, a spokesperson, tells the The Seattle Times. “How can the Seattle Police Department share terabytes of information we’re storing?”
— Ben Gilbert (@gilbert) December 19, 2014
Citizens only feel cameras increase accountability if they trust the devices are used properly, if they cannot be switched off at critical moments or if the video won’t be buried by scandal-averse commanders. But police departments cannot simply post raw video of every arrest to YouTube. To protect individuals’ privacy, state law prevents police from releasing details like the faces of juveniles or sexual assault victims as well medical details or mental health history, explains Mary Perry, the police department’s counsel.
But currently, removing a simple cut from a one-minute video “can take specialists upward of half an hour, whereas more complicated edits — like blurring multiple faces or pieces of audio — can take much, much longer,” an S.P.D. statement says. That’s a problem when the police are already burning an average of 7,000 DVDs every month and will have even more as body cams are rolled out for the entire force.
Technologies like image-recognition seem to be the police’s best bet for a quicker, cheaper way to systematically redact sensitive information. “Government agencies don’t jump out to me to be at the forefront of technology research,” says Simon Winder, head of Impressive Machines, a tech company focused on robotics, machine learning and recognition software. But with such huge tasks, cities are primed to adopt cutting-edge solutions. “There are so many ways we can yet use technology,” Seattle’s mayor Ed Murray responds. “We want to be the number one digital city.”
One of the recurring topics the hackers discussed was what to do when an algorithm makes an error in identifying a person or a frame of video, particularly because so many are shot in the dark of night or in the blur of pursuing a suspect. “The problem is you can’t just say ‘oops’ when you violate someone’s right to privacy,” says Brandon Arp, a software developer at Groupon who attended the hackathon. He proposed a “very conservative” system that hides more information from a clip than required by law but allows for a person to request a manual secondary review of individual redactions.
Ideas like this emerged over the five-hour brainstorming session (and free lunch) in the basement of police headquarters, prompting officials to predict they will become a national model. Officer Patrick Michaud says he was “blown away” by the hackathon. “Options came out of it, which is what we look for,” he tells The Seattle Times. “A different way to look for problems always works for us.”
— Seattle Police Dept. (@SeattlePD) December 19, 2014
DON’T MISS: This Is What Community Oriented Policing Looks Like