The Cybersecurity 202: Maryland scored a win using facial recognition software in Annapolis shooting

Monday , July 02, 2018 - 6:10 AM

Derek Hawkins

(c) 2018, The Washington Post.

Police responding to the mass shooting at the Capital Gazette newspaper in Annapolis faced a perfect storm of problems when they took the suspected gunman into custody: The man had no identification, he wouldn’t speak to investigators, and a fingerprint database wasn’t producing results.

But they had a backup plan: Investigators ran his photo in Maryland’s state-of-the-art facial recognition database. The system quickly returned a match.

The case appears to represent a highly successful deployment of the controversial technology, saving investigators critical time as they scrambled to identify a suspect and find out whether he was working alone. And it could boost arguments from law enforcement in favor of facial recognition at a time when systems such as Maryland’s have fallen under intense criticism from privacy advocates and civil rights groups who say they could be used to surveil innocent people or reinforce racial profiling. Proponents of the technology will surely point to it as a compelling example of the value such systems could offer police departments.

“This sensational case will probably awaken and create an awareness that will bring a lot more attention” to facial recognition systems among law enforcement agencies that haven’t adopted them, said Tom Joyce, a former lieutenant commander with the New York City Police Department’s cold case squad.

“It’s an extremely progressive idea, and it’s really important to the investigation to accelerate the case,” said Joyce, now vice president of Vigilant Solutions, which offers facial recognition services and other digital tools to law enforcement agencies. “You don’t know if he’s working alone. You don’t know if he’s got other victims. And you want to rapidly bring those to resolution. This was a great application of that.”

Maryland police were able to identify the man they say fatally shot five Capital Gazette staff members Thursday by feeding his picture into the Maryland Image Repository System, or MIRS, as my colleague Justin Jouvenal reported. The system uses algorithms to scan for a match across tens of millions of images from driver’s licenses, offender photos and mug shots from an FBI database.

Police officials said facial recognition was their best option for identifying the alleged shooter, Jarrod Ramos, after the fingerprint database was slow to return a hit. In a news conference the morning after the shooting, Anne Arundel County Police Chief Timothy Altomare acknowledged the state’s system had “come under fire from civil libertarians,” but said the investigation would otherwise have taken “much longer.”

“It was a huge win for us last night and thus for the citizens of Anne Arundel County,” Altomare told reporters.

“The facial recognition system performed as designed,” Stephen T. Moyer, secretary of Maryland’s Department of Public Safety and Correctional Services (DPSCS), told Justin in a statement. “It has been and continues to be a valuable tool for fighting crime in our state.”

But critics of the technology say not every case is as straightforward as this one. They also note that facial recognition systems tend to misidentify African Americans more often than whites and could allow police to conduct real-time surveillance against people not suspected of crimes.

“While the method seems to have performed well in Ramos’ case, there are still significant civil liberties concerns around the way police use MIRS,” wrote Russell Brandom of The Verge. “Police are supposed to remove people who were arrested but found innocent, but since the system is rarely audited, it’s hard to say if that’s actually happening. There are also racial justice concerns, given racial disparities in rates of arrests, compounded by higher error rates for African Americans in many facial recognition algorithms.”

In the wake of the shooting, Samuel Sinyangwe, a prominent racial justice activist and data scientist, pointed out there are few checks on how police use systems such as Maryland’s:

Samuel Sinyangwe tweeted “Note that half of all US adults are in facial recognition databases and there is very little oversight, testing for accuracy, or limits on how police use this software.”

And the Electronic Frontier Foundation, a digital rights organization, said that even flawless facial recognition would bring major privacy concerns, tweeting “Let’s say facial recognition improves-that it produces correct matches 100% of the time. Then what? Well, it means we can’t walk around ‘without the government knowing who we are, where we are, and who we’re talking to,‘ explains EFF’s Jen Lynch.”

Meanwhile, Clint Watts, a former FBI agent and senior fellow at George Washington University’s Center for Cyber and Homeland Security, wondered why police couldn’t have identified the suspect through other means, tweeting “Maybe there’s an investigative reason I’m not aware of, but if suspect has gone to this level of effort to hide his identity, facial recognition might be tougher than one thinks. Particularly if he’s not on social media.”

Concerns such as those have dominated an escalating debate over facial recognition that has unfolded in recent weeks.

In May, privacy groups and several House lawmakers called on Amazon to reconsider selling powerful facial recognition software at a bargain price to law enforcement, as I’ve reported. They said the service, known as Rekognition, could be used to inappropriately surveil innocent Americans and exacerbate racial profiling in black communities. Amid the pressure, the Orlando Police Department said last week it wouldn’t immediately renew a pilot program for the service, as my colleague Hazma Shaban reported.

Even a prominent facial recognition entrepreneur says he’s worried about abuse. In a widely shared commentary for TechCrunch last week, Brian Brackeen, the chief executive of the facial-recognition company Kairos, wrote that he wouldn’t sell his tools to police because the technology contained biases against people of color. He urged others to follow suit. “In a social climate wracked with protests and angst around disproportionate prison populations and police misconduct,” he wrote, “engaging software that is clearly not ready for civil use in law enforcement activities does not serve citizens, and will only lead to further unrest.”


Sign up for e-mail news updates.