The Surveillance Technology Oversight Project (STOP), a nonprofit hosted by the Urban Justice Center that is dedicated to advocating against state-sponsored surveillance, today filed suit with Debevoise & Plimpton against the New York Police Department (NYPD). The suit alleges the NYPD refused to disclose records from surveillance it used to track public sentiment about the department. STOP claims the lawsuit, which was filed in the New York State Supreme Court, comes over 10 months after the NYPD received STOP’s Freedom of Information Law request.
According to STOP, the NYPD has since 2016 awarded contracts totaling at least $3 million on an algorithm-powered “sentiment meter” that purports to measure attitudes toward the NYPD by precinct. Law enforcement officials jointly developed the sentiment meter with Elucd, a Brooklyn, New York-based startup with ties to former Police Commissioner Bill Bratton. Elucd reportedly collected data from over 250,000 New Yorkers — including 44,185 in 2020 alone. And the NYPD’s contract was extended in January for three years at a rate of $1.39 million per year.
Elucd uses targeted digital advertising to “send” surveys in English, Spanish, Chinese, and Russian to users on a range of websites and social media platforms, through traditional landline calls, and over 50,000 smartphone apps — including Candy Crush and WeatherBug. The company draws on U.S. Census American Community Service data for statistics on locations and adjusts recipients of its survey to obtain representative samples. It asks survey-takers to respond to 10 rotating questions — like “Do the police in my neighborhood treat local residents with respect?” and “How safe do you feel in your neighborhood?” — using a 1 to 10 scale.
The sentiment meter divides New York City into 297 individual precinct sectors and displays the public’s perception of police per precinct. The color green indicates a positive change in perception, while red signals a negative shift. Approximately 10,000 people complete Elucd’s survey in New York City each month, according to Elucd cofounder and CEO Michael Simon, who headed former President Barack Obama’s data analytics team during the 2008 campaign. But STOP alleges that the NYPD has only sparingly released data on its trust ratings (which are aggregated monthly); has consistently declined to release data at the neighborhood or precinct level; and hasn’t disclosed how changes in sentiment are calculated and what adjustments the NYPD makes in its policing practices as a result.
In December 2019, the New York Post revealed that the NYPD’s overall sentiment meter trust rating improved from an initial 6.1 in September 2016 to 6.6 in August 2019, while its safety score dipped from 6.6 to 6.3. The paper further reported that internal usage of the sentiment meter among NYPD officers was inconsistent, but as STOP’s petition before the New York Supreme Court notes, algorithms like those underpinning the sentiment meter have the potential to exhibit bias against vulnerable demographic groups, such as minorities, women, and the disabled.
Moreover, the data NYPD seeks to gather could be integrated or consolidated within existing databases, creating a window into citizens’ lives. (The NYPD hasn’t revealed whether individuals can be identified from the data the surveys collect, nor has it shared any other cities, states, or federal agencies the data is shared with.)
The NYPD’s refusal to disclose information about the sentiment meter could run afoul of the Public Oversight of Surveillance Technology (POST) Act, a bill that requires the department to disclose its use of surveillance technologies and develop policies to govern its deployment of those tools. The NYPD maintains that under its renewed contract it doesn’t have access to or ownership of Elucd’s raw data — including respondents’ age, gender, race, and location — only its own analysis of that data.
“It’s outrageous that the NYPD spent years building software to track New Yorkers’ sentiment on policing,” STOP executive director Albert Fox Cahn said in a statement. “Instead of developing this invasive software, the NYPD could have just listened to the countless New Yorkers who demonstrated to defund the police. It’s Orwellian to think of the NYPD building out a map, block by block, of how the public views police. And it’s indefensible for the NYPD to try to hide records about this mass surveillance from the public.”
Elucd, which claims to have a budget in the multimillion-dollar range, including a combined $1 million from Y Combinator and the Omidyar Network and $289,000 from the New York City Police Foundation, says it has worked with over 15 police departments and governments. But it has a spotty record when it comes to transparency. According to the Marshall Project, the Research and Evaluation Center at John Jay College of Criminal Justice shaved off nearly a third of Elucd’s $1.4 million grant and hired a think tank to dig into the company’s data science. The National Opinion Research Center issued 18 recommended changes, and while Elucd adopted a few of the suggestions, it characterized most of the feedback as “outdated.”
But Elucd claims success stories in other cities, like Chicago and Los Angeles, where local police have been measuring public opinion with the company’s surveys since 2018. In Grand Rapids, Michigan — which used drug forfeiture funds to pay Elucd $150,000 for two years of service — law enforcement reportedly assigned more foot, bike, and horse patrols to neighborhoods with negative perceptions of police. Whether this represents a win for the public is a subject of heated debate.
“The relationship between police and the communities they serve is at a crisis in many cities,” Simon told the Marshall Report in a July 2018 interview. “And they have no way of measuring what success is right now.”