27 Mar Privacy or Safety? Experts Call Surveillance Debate Misleading

“Cameras don’t stop crime. They just record them,” said Rebecca Gerny during a discussion surveillance. Later, Aaron Peskin said, “Facial recognition technology disproportionately misidentified people of color because the algorithm was created by white dudes.” (Danielle Parenteau-Decker / The CC Pulse)

By Danielle Parenteau-Decker 

The debate over surveillance is often framed as one between privacy and safety. But a discussion hosted recently by American Community Media suggests that is misleading, as growing concerns of racial profiling and government overreach point to surveillance as being potentially more dangerous than the harms it claims to prevent. 

The conversation, held Feb. 27 at the World Affairs Council in San Francisco, examined the debate. Moderator Jaya Padmanabhan, executive director of ACoM, said there is a “false two-way narrative” around whether Flock Safety provides safety or surveillance, summing up the issue as a whole. 

In Richmond, the debate is playing out in real time.  

Last year, the Richmond Police Department turned off its Flock license plate readers after learning that federal agencies could potentially access the data. This month, the Richmond City Council voted 4-3 to turn them back on. 

Those opposed said Flock was untrustworthy and that the system endangers immigrants under the guise of public safety. 

The chief of police told the council that vehicle thefts surged with the cameras turned off, but a council member pointed out another period in which vehicle thefts increased while they were active. 

“Cameras don’t stop crime. They just record them,” said Rebecca Gerney of East Bay Sanctuary Covenant at the Feb. 27 event. 

Similarly, Ariel G. Ruiz Soto, a senior policy analyst at the Migration Policy Institute, said that “more data [gained from surveillance] has not automatically translated into more deportations despite the Trump administration’s focus.” 

Rather, he said, it has created “chilling effects including chilling access for health care, schools, courts and public benefits, including potentially undermining tax compliance and other core government functions.” 

Sources of that data include “very common things that many of us do not think often about like DMV records,” he said. 

Ruiz Soto also said the government wants to collect biometric data from immigrants and their sponsors and family members who are U.S. citizens, warning that such systems can expand beyond their initial targets. 

That biometric data could even include your heartbeat. 

Jacob Ward, a reporter and former editor in chief of Popular Science magazine, said there is technology to identify your heartbeat that is more accurate than facial recognition. 

And, naturally, harder to evade. 

“You can’t leave your heart at home. You can’t cover it with a mask,” he said. 

Ward added that facial recognition is less accurate at identifying Black, Brown and female faces “as it is identifying faces that look like mine.” 

To Aaron Peskin, a former San Francisco supervisor, that is no surprise. 

“These algorithms reflect the biases of the programmers who create them who tend to overwhelmingly be white men,” he said. “So guess what? Facial recognition technology disproportionately misidentified people of color because the algorithm was created by white dudes.” 

Jacob Snow, a senior staff attorney for the American Civil Liberties Union of Northern California, said the nonprofit revealed that Amazon was selling facial recognition software to law enforcement, then put that software to the test. 

In 2018, ACLU NorCal used Amazon’s Rekognition tool to compare around 20,000 publicly available mugshots with photos of all the then-current members of Congress. It found 28 matches — “All of those, to be clear, were false matches,” Snow said — most of whom were people of color, including civil rights activist John Lewis and five others in the Congressional Black Caucus. A similar test the following year wrongly matched 26 California lawmakers with arrest photos. 

“The dangers are both when it’s flawed and when it works,” Snow said, specifying that it could be used to locate “somebody in a medical clinic who’s seeking abortion” or “an undocumented person trying to go about their daily lives.” 

Ruiz Soto said surveillance used in immigration enforcement doesn’t always go after “the worst of the worst.”

Still, it can seem to some that the only people who need to be worried about surveillance are those who have reason to worry that law enforcement might be after them. 

Juan Sebastian Pinto, who writes about technology and civil rights, challenged that idea, saying companies are not just collecting individual data but using it to predict behavior. 

He said that artificial intelligence companies “don’t necessarily want your specific information … but they want to learn about everything” that will help them predict how “everyone who shares your characteristics” will behave in the future. 

Peskin said there is “a whole host of technologies” that can be used to track people and behavior, a lot of which “you might not, and I didn’t think of, as surveillance technology [that] permeates so many aspects of our lives.” 

And tech companies can get the data they want from things that are made to seem not invasive but helpful or even fun. 

Ward pointed to the 10-year social media challenge in which someone posts two photos of themselves 10 years apart as an “amazing way” to train facial recognition software. 

“Persistence over the course of a lifespan is one of the hardest things to do in facial recognition,” he said. “But they’ve solved that for the most part, in part, because we fed them all the information” needed to have that happen. 

Speaking about devices like Alexa, Ward said that even an intelligence chief for the Stasi — the notorious East German secret police known for mass surveillance — “could never have imagined a world in which we actually paid money to put a mic in our house, and yet that’s the world we live in.” 

But as widespread as this technology may be, people are finding ways to push back.  

Pinto said resistance to surveillance companies is being led largely by women, while others, including young people, are beginning to reject smartphones and social media 

Ward also said one word gives him hope: “clanker,” slang for someone who uses A.I. too much — as in “Don’t be a clanker.”