The Washington PostDemocracy Dies in Darkness

America is awash in cameras, a double-edged sword for protesters and police

Smartphone cameras, home security cameras, traffic cameras — digital eyes are a boon and danger to protesters

Police and businesses have sophisticated surveillance tech, and social media posts from a protest can lead to harassment. Protect your privacy with these steps. (Video: The Washington Post)

On Saturday night, as protests were still taking place in city streets across the country, the Dallas police department put out a call for help on Twitter. It asked anyone who had video from the protests showing “illegal activity” to upload it to its anonymous tip app, iWatch Dallas.

What it got was a different kind of protest, in the form of a flood of videos and images of K-pop stars performing. The department later tweeted that the app was down because of technical difficulties.

In the tense and escalating standoffs between law enforcement and protesters that have now spread to more than 100 cities in the wake of George Floyd’s killing in Minneapolis police custody, photos and video footage are being collected and wielded by all sides. And there is no shortage of cameras to pull from. Law enforcement agencies have existing networks of surveillance cameras and body cameras worn by officers, as well as face- and object-recognition software. Large retail and food chains have similar security systems and traditionally will share footage with police if it is part of an investigation. Protesters and journalists shoot their own videos, the latter often on smartphones or small cameras such as GoPros.

Then there’s the vast unofficial network of personal security cameras, including Amazon’s Ring and Google’s Nest cameras, that make it easy to record people and gatherings outside homes or some small businesses. Even Tesla and other vehicles’ cameras can provide recordings of their surroundings. There’s no need to even watch all the videos — many work with apps that automatically detect movement or people. Tools such as Ring’s Neighbors app make it easy for police to ask for footage from people with cameras in specific areas and easy for people to then share it with law enforcement.

But it was also cameras that helped set off the latest round of protests over police brutality. Without them, Floyd’s death would not have been captured from multiple angles and shared so widely.

The Washington Post reconstructed some of the events preceding the death of George Floyd using security footage, EMS audio and cellphone video. (Video: The Washington Post)

The death of George Floyd: What video and other records show about his final minutes

“The outrage of seeing a murder has created a national conversation, whereas some of the other police shootings haven’t,” said Andrew Guthrie Ferguson, a law professor at American University and the author of “The Rise of Big Data Policing.”

Cameras have been present at protests for years, but more recently, the technology that allows people to share footage or analyze it has changed significantly. With the growth of tech companies and their offerings have come increased concerns regarding how and when people are being watched — as well as how that information may be used by law enforcement and the government. Adding to these concerns is the growing use of facial-recognition software and other tools powered by artificial intelligence that are used by law enforcement to track down information and suspects.

Before Ayo Omolewa attended a protest in the District on Saturday, a friend texted him to put his phone in airplane mode and not use his GPS so as to avoid being tracked by law enforcement. The Maryland-based cloud-security engineer said he stays aware of where cameras are that could be used to identify him.

Federal study confirms racial bias of many facial-recognition systems

He’s less concerned about ever being caught on camera and more worried about the possibility that he could be put in danger by being identified at a protest.

“Surveillance is a really big issue right now, and it really shouldn’t be,” he said.

It’s no longer just a tool of the police. Smartphone camera footage of crowds of protesters peacefully marching and resisting police officers’ advancements has spread on social media, bolstering supporters’ cry that most people are gathering nonviolently. Phone cameras also record scenes that become rallying cries for protesters, including the moment when a police car drove into a crowd of protesters in Brooklyn on Saturday. Some videos from protests have led to disciplinary actions against officers, including firings.

Live-streaming from smartphones has become a common feature on most social media platforms, including Facebook, Instagram, YouTube and Twitter’s Periscope, and protesters are able to show what is happening in real time and capture incidents of police brutality that might have been erased in the past. On Monday night, a group of protesters took refuge in a stranger’s house in Washington after the city’s curfew, leading to an all-night standoff with police that was live-streamed and went viral.

“The ability for the public to document what is going on is an important tool for holding powerful people and institutions accountable, including the police,” said Evan Greer, deputy director of the digital advocacy group Fight for the Future. “The availability of, particularly, smartphone cameras has dramatically increased the number of instances that we see.”

D.C. resident Rahul Dubey gave refuge to around 60 protesters inside his home after law enforcement had pushed them down his street. (Video: The Washington Post, Photo: Bonnie Jo Mount/The Washington Post)

But all those cameras — and the devices they’re attached to — also give law enforcement more awareness of the location and identity of protesters, offering unique identifiers that many protesters try to avoid. Videos and photos shared online often carry metadata that provides device and location information, something Greer says people can avoid sharing by taking a screenshot of a photo and sharing that instead.

Oregon became a testing ground for Amazon’s facial-recognition policing. But what if Rekognition gets it wrong?

Facial-recognition technology is another concern. The tools have become more commonly used as law enforcement agencies purchase face-detection software, including Amazon’s Rekognition and Clearview AI. They are typically used to match sometimes low-quality images of faces grabbed from security cameras, social media or smartphone photos with vast databases of mug shots or images scraped from the Internet to find a person’s name and contact information. Civil rights groups and AI researchers have said the systems are often inaccurate, especially when used to identify women and people of color. A federal study last year also found a high incidence of bias in the software.

Some protesters have asked that footage showing faces without consent not be widely circulated on social media or by journalists because police have the same access to the videos as everyone else. Some police departments have tools that scrape sites like Facebook, YouTube and Twitter for relevant videos. Wearing masks, like the ones currently recommended to limit the spread of the novel coronavirus, can make it harder for facial-recognition tools to identify people but not impossible.

In Chicago, protester Brian Grieshaber and his brother were careful to wear masks — especially important during the pandemic, he said — and hats, sunglasses and generic clothing to avoid being identified.

Use of medical helicopters to target protesters a ‘foolish’ violation of norms, experts say

“We have been learning tactics from the protests in Hong Kong,” Grieshaber said.

Law enforcement agencies do not regularly disclose how they investigate people, including whether tools like Rekognition are used.

Amazon did not immediately return a request for comment about Rekognition. (Amazon chief executive Jeff Bezos owns The Washington Post.) Clearview AI did not immediately respond to a request for comment.

A sheriff's office in Oregon is using Amazon Rekognition to help track down criminal suspects. Is this a dangerous precedent against privacy? (Video: Drew Harwell, Jhaan Elker/The Washington Post)

People and businesses with their own security systems are a coveted part of growing surveillance networks. At least 1,350 law enforcement agencies are on Amazon-owned home security camera company Ring’s Neighbors app. It lets police see videos Ring camera users post to the app, as well as put out requests for relevant footage from users in specific areas.

Public safety agencies have access only if a user posts it or shares it, said Ring’s head of communications, Yassi Shahmiri. “Further, all video requests are facilitated by Ring, meaning public safety agencies cannot see user names, contact information, video content or location unless a user voluntarily chooses to share it.”

Ring and Nest helped normalize American surveillance and turned us into a nation of voyeurs

Separate from the social network apps, governments and law enforcement agencies can also submit official legal requests directly to companies for information on users. For example, a police department might ask Google for video footage from a user’s Nest camera stored on its servers. Nest accounts allow footage to be stored in Google’s cloud for a set period of time. The companies all have their own policies to help decide which are approved.

“Google does not have partnerships with law enforcement regarding the sharing or disclosure of user data, including Nest camera footage,” said Google spokesman Evan Barbour Grippi. “We have a formal process that we have long followed to appropriately vet requests for data users store with us.”

Additionally, many police departments have camera registry programs that skip the tech middlemen. People are asked to tell the city that they have a private security camera, where they’re located, and how to contact them. The programs are used by police departments to put out calls for footage to specific people they have on file in areas they’re investigating.

Gabriel Meyer-Lee protested on Saturday in Burlington, Vt. Meyer-Lee, a PhD student studying computer science, said they weren’t overly concerned about the potential use of facial-recognition software in the small city, but other protesters may worry about it in other, larger cities. Fellow protesters work to protect identities by scrubbing metadata off photos or obscuring faces in the photos they share.

Retailers and restaurants across the U.S. close their doors amid protests

Taking note of the multitude of cameras that may be watching at any time has become more a part of conversation generally, they said. “People haven’t fully adapted to it but they are becoming more aware.”

Some cameras can also provide some much-needed peace of mind. Anna Ruth Williams, the founder of public relations firm ARPR in Atlanta, said she has been watching the live footage each night from the four Nest cameras at her business in the Atlantic Station shopping area in Atlanta. The area was a potential target for looting, property managers warned, so she secured valuables.

Williams, who supports the peaceful protests, said she believes the looters in her area are mostly from outside, unaffiliated groups. So far, her office has not been affected.

“I’m kind of living in this world of personal pain and sadness,” she said. “But I’m also having to be a business owner that has to be protective of my business property and my people in the face of outside groups.”