With help from Ella Creamer, Jesse Naranjo, Rishika Dugyala and Teresa Wiltz
Recast family, an important update to share with you, our loyal readers.
Starting today, The Recast will now be a weekly publication, rather than twice weekly. With the 2024 election rapidly approaching and Brakkton’s increased role in reporting on the campaign, we’re scaling back newsletter production to focus on coverage across POLITICO’s many platforms. Don’t worry: We will still publish The Recast every Tuesday, complete with our unique reporting and analysis.
As always, if you have any suggestions on what we should be digging into or who we should be interviewing, drop us a line at [email protected].
All the best,
The Recast Team
Facial recognition technology is omnipresent in our daily lives — and it shows no signs of abating.
It’s used to open our smartphones, to diagnose diseases more easily and to track attendance at public events. It’s so ubiquitous, we often don’t think about it — or how much it’s already impacted our lives.
But the area where the tool elicits the most visceral response is when it’s used by law enforcement agencies.
Civil rights organizations such as the NAACP and the ACLU denounce its prevalence, arguing the technology is dangerous and violates Americans’ civil liberties while also perpetuating bias in law enforcement against communities of color. Others contend it is a necessary and effective tool that police need to stem the growing tide of violent crime, particularly in urban areas.
Today, POLITICO is reporting on the New Orleans Police Department’s use of facial recognition technology. Why New Orleans? Last year, faced with an uptick in violent crime, the City Council voted to allow police to use facial-recognition software to identify suspects — overturning its own ban on the technology, initially put in place in late 2020 following the death of George Floyd and the nationwide protests his murder ignited.
But now the effectiveness of the technology is being questioned.
NOPD officers requested its use just 15 times since facial recognition technology was restored last October. Statistics indicate that nine failed to make a match. Among the six that did return matches, three turned out to be flat-out wrong. Two are pending investigation. And just one has led to an arrest.
Here’s another sobering stat: Of those 15 requests, just one was for a white suspect.
Was The Recast forwarded to you by a friend? Don’t forget to subscribe to the newsletter here.
You’ll get a weekly breakdown of how race and identity are the DNA of American politics and policy.
New Orleans City Council Member Eugene Green has no qualms with that. He sponsored the legislation that rolled back the city’s surveillance ban, paving the way for the NOPD to use facial recognition technology when investigating violent crimes. (Last year, New Orleans ranked first in the nation for homicide rates.)
When the use of facial recognition technology was reinstated, city lawmakers mandated it could only be used to investigate violent crimes and that a supervisor must sign off on any officer requesting its use. New Orleans itself does not have its own database; instead the city accesses the technology through a partnership it has with a state-run program called the Louisiana State Analytical and Fusion Exchange.
Green, who is Black and represents a majority-Black district, says despite the low arrest figures since the technology was reinstated, the city’s police should be commended. As he sees it, critics who claimed police would use it in a discriminatory fashion have been proven wrong.
We chop it up about why he believes New Orleans has the blueprint for tracking and implementing safeguards for its use — and why he says the technology is needed, just in case there’s a Jan. 6-like event in his city.
◆◆◆
This interview has been edited for length and clarity.
THE RECAST: Before we jump into some of the findings of the recent report, can you walk me through what led to you introducing a bill to repeal the city’s facial recognition ban?
GREEN: When you say it was banned, we don’t have the ability to influence the state. For example, which state troopers have been able to use it; the federal government has been able to use it.
What was banned was the New Orleans Police Department accessing the technology from others. That’s very, very, very important. This legislation that I passed allowed the police department to access that technology from places that they had previously been banned from accessing before.
THE RECAST: So for my clarification, you are saying the New Orleans PD could not access tools that other agencies in and around the city could. And what your bill did was lift that shield for NOPD?
GREEN: That’s absolutely right. Well, access to the technology that is already available. It’s important to use the word “access” because when the police department wants to use facial recognition, they have to go to another agency, and ask them to access what they have on the cameras.
THE RECAST: As my colleague, Alfred Ng reported, of the 15 facial recognition requests that actually went through, here are some stats: Documents indicate that nine of them failed to make a match. Three turned out to be wrong altogether. Two are pending investigation and only one has been arrested. This doesn’t sound as if it’s particularly effective.
GREEN: What are you saying that I said about effectiveness? It’s a tool.
I thought what the issue was [before NOPD had access to the technology] was that the police department was going to discriminate and find a reason to go after Black males. They were going to make assumptions and they were going to imprison them.
Well, you cannot hide the fact that it hasn’t. … That is a success for the police department. It shows that with a Black police chief [who retired in December], a Black district attorney, a Black mayor, Black head of agencies, such as the city attorney’s office, that they don’t have any interest in using the facial recognition technology to discriminate against Black males.
But it’s not something that should be banned. Because you just never know. You never know if you are going to have an attempted takeover of the government like Jan. 6, where they’re ending up using facial recognition to identify most of those who have been arrested. That’s the problem with having these type restrictions that were put in place previously.
What’s the point? Yes, it could be used in a discriminatory manner. But what the police have shown over the last year-plus is that they don’t intend to do that.
THE RECAST: But why continue using facial technology if it is not boosting arrests or reducing crime?
GREEN: We had 265 murders last year; not very many of them were committed by persons other than those who happened to be African American. And the crime is concentrated in what happens to be, in many instances, majority-Black communities. That can be a product of discrimination. That’s a product of racial separation, Jim Crow, that’s a product that goes back to slavery. It’s a product that goes back to shunting people into certain neighborhoods.
If we’re going to deal with the symptoms and deal with the reality, we have to deal with the crime numbers, but we have to look at who is committing the crime. There’s nobody alleging the city of New Orleans is using this in a discriminatory manner.
Why would we ever give up a tool that no one can say has been abused? It has not been.
I assert to you that the police department doesn’t have a tremendous need for its use.
[But] you should not throw out a tool unless you have a reason to throw it out. We don’t have a reason to throw it out. And I think that the police need to be praised for the lack of anybody being able to suggest that they abuse such a viable technology.
THE RECAST: When the New Orleans City Council moved to restore the use of facial recognition technology, it put in place some safeguards including tracking which officer made the request; what supervisor signed off on the request; and the kind of crime that was committed
Do you feel New Orleans has a blueprint that other jurisdictions can and should adopt?
GREEN: Of course! We’re under a consent decree.
We are in a position where the police make thousands of arrests a year and [few say] that the police are doing an unconstitutional job in making those arrests.
[There’s the] Public Integrity Bureau that watches police actions that the public can complain to. We also have the inspector general’s office which can initiate an investigation at the request of citizens. We also have a district attorney who can refuse charges. We also have a city attorney.
That’s four different layers of control in addition to the ordinance itself, which makes it clear that you have to identify who the officer is, and why they’re making the request. That’s why at the end of the day, you are only able to point to 15 instances where they simply asked if the images that they had matched under a facial recognition scenario.
TECH ON TRIAL
Rare testimony from a 2019 criminal trial also highlights how human bias still slips into a process many people think is automated and accurate, reports POLITICO’s Alfred Ng.
New Orleans public defender Walker Rick was in the middle of an armed robbery trial when he learned a human ultimately picked out his client based on a facial recognition match.
Police used surveillance footage from one of the many cameras that blanket New Orleans, and sent the image to the Louisiana State Analytics and Fusion Exchange, a state-operated surveillance network that conducts facial recognition searches. The staffer, who ran the software testified at trial that she got nearly 200 matches back, which is average for a facial recognition search. Rick’s client was in the first 10 results.
The staffer, who had worked with facial recognition for about two years after a two-day training session from the software provider, testified there were no scores or confidence rankings. She said the best matches tend to be on the first page, similar to a Google search result.
She ultimately picked out Rick’s client herself. But she didn’t single him out based on the facial recognition program — she made a judgment call based on his criminal history, which included a prior arrest for armed robbery, according to her testimony.
“We just kind of look at and decide if we think it could be a possible lead, and then we convey that it could be possible,” the staffer said, according to a transcript of the testimony.
The defendant ended up pleading guilty.
Lt. Melissa Matey, the Louisiana State Police’s statewide operations supervisor for public affairs declined comment.
Up until this testimony, Rick said he was under the impression that the state’s facial recognition software was an accurate, sophisticated algorithm. Instead, what he saw was a person with potential implicit bias behind the curtain operating a machine.
“She just flipped through and did a photo comparison and picked herself which person she believed looked most like the person,” Rick told POLITICO.
ICYMI @ POLITICO
Nikki Haley Rising — Check out the latest on GOP presidential candidate Nikki Haley. Her campaign is on the rise, even if she holds campaign stops at Iowa mortuaries. More from POLITICO’s Natalie Allison.
Isn’t This Some Bill Sh-t — The Biden administration is out with its latest push to ease the burden on Americans with student loans, POLITICO’s Michael Stratford reports. The latest hang up? The Education Department says a company it used to process payments didn’t send out bills to some 2.5 million users, resulting in missed payments.
Andy Kim Against the (New Jersey) Machine — Earlier this month, Rep. Andy Kim told us he knows he’s not central casting as he makes an insurgent run at the New Jersey Senate seat. POLITICO’s Brakkton Booker digs deeper, seeing what hurdles remain for him to win the Democratic nomination… and how he hopes to galvanize the state’s large AAPI voter population.
THE RECAST RECOMMENDS
Vauhini Vara’s story collection, “This Is Salvaged,” is an intimate, inventive close-up of relationships between family, friends, lovers and neighbors.
Posthumous secrets are revealed in the much-anticipated, Oprah-produced mysterious family drama “Black Cake,” set across multiple continents and generations.
Le Sserafim drops their first English-language track, “Perfect Night” — a collab with video game Overwatch 2. The MV has a fun dual narrative: The K-pop group are in dance rehearsals while animated Overwatch characters break traffic rules trying to make it to the group’s show on time.
“All Dirt Roads Taste of Salt,” out Friday, chronicles the life of one Black woman in Mississippi through tender vignettes and a meditative soundscape.
Kid Cudi can’t stop smiling as he dances in a Seoul market for “Ill What I Bleed.”
TikTok of the Day: Happy Halloween!
