“Is NYC’s new gunshot detection system recording private conversations?” asks Fusion in a recent story about ShotSpotter, a sensor technology currently being set up in the Bronx and Brooklyn.
ShotSpotter sensors use microphone and satellite technology to detect, locate and report gunshots to police. Critics worry that the microphones are prone to false alarms, and more troubling, appear to vacuum up street-level conversations in the neighborhoods where they have been installed. Evidence from conversations recorded by ShotSpotter microphones has been used to prosecute criminals in court.
While questions linger for watchdog and privacy groups about the use of ShotSpotter technology, an aggressive lobbying campaign has helped ensure the devices have been deployed in over 90 cities across the country.
The Ferguson Group, a Washington, D.C.-based lobbying firm, boasts that it secured more than $7 million in federal funding to support the purchase of ShotSpotter. “TFG has conversations with interested communities and discusses process and assesses viability of request [sic], drafts and provides briefing sheets to communities and submits requests to their House and Senate delegation,” reads a case study posted on The Ferguson Group’s website.
ShotSpotter contracts with four D.C. lobbying shops, including the powerhouse Squire Patton Boggs and the Raben Group, the firm that helps orchestrate Mayors Against Illegal Guns, an advocacy group closely aligned with former New York mayor Michael Bloomberg and various police unions across the country. The firm also has an array of local and state lobbyists on contract. In New York City, for instance, the company retained Greenberg Traurig in the past, and now works with a former aides to Sheldon Silver and Bloomberg through the firm Mercury Public Affairs.
The company’s approach is detailed in emails from Phil Dailly, southeast region sales director for ShotSpotter, to the City of Miami. Dailly references a supportive city resolution and lists viable funding mechanisms, including purchasing the technology through the Community Oriented Policing program, a special fund administered by the Department of Justice, or through police department asset forfeiture money, funds often raised through drug busts. Promotional materials also list the DOJ’s Justice Assistance Grant program, Public Housing Agencies and Community Benefit Funds as potential funding sources. The company retained two local lobbyists in Miami to help move the process along.
The company also maintains close ties with leading law enforcement officials. ShotSpotter’s senior vice president David Chipman is a former senior official at the Bureau of Alcohol, Tobacco, Firearms and Explosives and former fellow to the International Association of Chiefs of Police. Before returning to the New York Police Department as police commissioner, William J. Bratton served as a board member to ShotSpotter. (Bratton said he recused himself from the NYPD’s decision to embark on a pilot program in New York City this year.)
The company has downplayed privacy concerns. ShotSpotter vice president Lydia Barrett, asked about conversations recorded by the technology submitted as evidence in court, told South Coast Today that it is a “very unusual circumstance if (the sensors) actually picked up any voices,” adding that, “It’s an acoustic sensor. It’s not a microphone, and it’s only activated when a loud boom or bang happens.”
However, a WNYC investigation in 2013 found that 75 percent of the incidents reported by the company were false alarms, alerts in which audio recordings were made in which there was likely no crime in progress. ShotSpotter’s own privacy policy explains that it is constantly recording in order to be able to provide police with audio beginning two seconds before a gunshot and ending four seconds after.
ShotSpotter’s privacy policy claims this audio is “erased and overwritten” and “lost permanently” if its system does not sense a gunshot. However, even if this is true, the policy also states that ShotSpotter has detected and recorded “3 million incidents” over the past 10 years. This also indicates the sensors report a staggering level of false alarms, and that the company has permanently recorded 18 million seconds — in other words, 5,000 hours or approximately seven months — of audio. According to a promotional document emailed to Miami city officials by ShotSpotter’s sales team, the technology allows end users to retain this audio online for two years and offline for another five.
Photo: John Moore/Getty
Ok, so it’s designed to capture gun bursts but picks up audio of conversations that’ve been used in court? To have the device recording all the time to pickup gunshots is kind of a stretch don’t ya’ think? I see an intentional repurposing (or original intent) of the technology taking hold. It’s illegal to record voice conversations-so this is an end-run around privacy.
There’s not very much wrong with this system. See the Washington Post article for a non-hysterical approach to it’s values and shortcomings.
We need a machine to fix the machines which protect us from machines we use to protect us from machines.
For a far less hysterical report on ShotSpotter, see this article from the Washington Post. There’s value to the system. It’s not perfect and it is of no use for listening to conversations. How many people sit up on poles 3 stories about the ground? Sometimes our authors make reaches to find controversy.
If for some reason the link is not printed, just google shotspotter site:washingtonpost.com
http://www.washingtonpost.com/investigations/shotspotter-detection-system-documents-39000-shooting-incidents-in-the-district/2013/11/02/055f8e9c-2ab1-11e3-8ade-a1f23cda135e_story.html
I’ll have to agree with Benito and Mike: Microphones are not human ears, therefor they don’t pick up what our ears pick up. An omni-directional mic (aka omni-directional acoustic sensor) is not suitable for recording a conversation on a street, even when people scream, unless it is right next to the people talking loudly/shouting/screaming. But placing mics at that height is not be useful if you want to cover a whole area. This still leaves the question unanswered, how the snippets used in court where recorded. The possibilities are endless, from cell phone mics to surveillance cameras, we are surrounded by sensors.
The other very interesting question concerning ShotSpotter is, why a system, that reportedly produces 75% false alarms has been deployed in 90 cities across the country and whether anyone has checked its efficiency and compared it to other alarm systems. After all DoJ tax money is going into this.
I remember suggesting such a system in the 90s. After all, a gunshot is not a private matter; sound physically penetrates our abodes, and becomes legitimate for us to listen to. Any kid with a videocamera can point it out his window, and if your conversation on the street below is audible, it’s hard to say he doesn’t have the right to record from his own house. But the details are more important now, and we should indeed still be on guard against deliberate expenditure to create surveillance infrastructure.
If the system indeed has a self-contained little box that stores a few seconds of audio on its little flash drive and only transmits them to document a gunshot, then I see little real spying potential. But if it “records” data by streaming it over the internet to a central server, which has a “policy” of deleting unused audio, then that is something else again. Or if a potential gunshot is always being registered, 24 hours a day.
Still, the most telling difference is the same now as it was then: how high are the microphones? To get an accurate triangulation of a gunshot’s location, they should be as high up as possible, as far as possible from buildings that confuse the detector with echoes. But of course to spy on people having conversations, they should be at about mouth level on well-used street corners. It shouldn’t be hard to sniff out what the designers really want by looking at their behavior.
The earliest patent application is 1994 (http://www.shotspotter.com/patents/patent/5-504-717-system-for-effective-control-of-urban-environment-security). OK, not quite the 30 years I said below. You are right, picking up gunshots and zooming in on conversations are two different things. I think there are more important and interesting things to be concerned with. Almost all video cameras (and how many of them are there?) are situated to be better conversation recorders, and there is no reason why they cannot record conversations as well as their video function with no one the wiser.
Fusion link points back to TI. Am I the only one who looks at original articles? :(
No, but that did not seem like a major problem with the article.
–George Orwell, 1984
Detecting an exercise of 2nd Amendment rights! How dare they?
I know, the people here only believe in about half of the “Bill of Rights” and really hate the Constitution since it limits their ability to engineer a society.
But consider if they were walking “drug dogs” through every neighborhood, along every street, and had them sniff every person (or detain them when they “indicated”).
Ought we not drop both the war on drugs and the war on guns?
(I’m waiting until someone finds the cheapest firecracker that will trigger the system).
Shotspotter? “…and more troubling, appear to vacuum up street-level conversations in the neighborhoods where it has been installed” Seriously? Microphones with broad directional patterns designed to pick up over a wide area can zoom in on conversations? I do not think so. If you are yelling loud enough for shot-spotter to pick up, privacy is not your goal.
And wow! Seven months of audio! All in little six second intervals where only the loudest sounds can be picked up.
This is the dumbest article I have read in months.
Mike S. –
I didn’t find this article dumb at all. That South Coast today link was quite interesting. One person interviewed is attorney Richard Cole, vice chairman of the Massachusetts Bar Association’s Individual Rights and Responsibilities Section (at the time, anyway). He says (from the article:) But the fact that ShotSpotter, with its primary function meant to be gunshot detection, can apparently record audio raises other questions, Cole said.
“Can (ShotSpotter) pick up private conversations? That is the broader issue that this case raises,” Cole said.
Also from that article: “If that is what the system does, then it does raise concerns. We’d just like to know more about what the technical aspects are,” said Christopher Ott, communications director for the American Civil Liberties Union of Massachusetts.
It is indeed the unknowns about this that trouble me. And the propensity for TPTB to skirt around things like oversight and transparency.
Shotspotter has been around a long time, more than 30 years, I believe. Why the interest now? Why the questions about how it works? Although the algorithms for correlating the information from different microphones are proprietary, it is otherwise not very complicated. Just pick up loud sounds located not very close to your microphones (because you do not have that many). This is a horrible way to go after normal level conversations. If you want to pick up a conversation that is not close in a noisy environment, you need to use a very directional microphone, aimed at a particular spot. No good for picking gun shots over a wide area. (If the speaker is very close to a microphone you can use noise cancelation techniques to improve reception, but that does not work here.) So we have a technique that some people say does not do its intended job all that well. How well is it going to do something else that it is poorly designed for?
I had never heard of this before – and you say it’s been around what – more than 30 years? Hmmm.
You ask, why the interest now? Maybe there’s more of an interest about such things now – at least in some quarters. I also forgot to mention the fact that this system doesn’t seem to be very good at its supposed job, something you picked up on. That alone makes me wonder why this is used at all.
Anyway, I’ve grown to be very skeptical about “official” explanations of anything. You can’t trust that things are as they seem. So I’d be wary of this system for lots of reasons.
Thanks for drawing attention to this.
“We know that the NSA can listen to all we say if we are near enough to a device it can turn on. (Quick: How close are you as you read this to an electronic device that the NSA can access and use as a listening device?) And we also know that the feds gave secret roadside listening devices to about 50 local police departments, which acquired them generally without the public consent of elected officials in return for oaths not to reveal the source of the hardware. It came from the secret budget of the CIA, which is prohibited by law from spying in the US. What’s going on here? What’s going on here is government’s fixation on spying and lying.” (Andrew P. Napolitano)
More here:
http://original.antiwar.com/andrew-p-napolitano/2015/03/25/the-governments-fixation-on-spying-and-lying/
The Big Question to ask here would be – How many electronic gadgets do you possess that is not “Made in China”?
I find it difficult to find even one. If the Chinese have the technology to secretly route all communication to their servers then it’s a very dangerous world that we live in.
Thanks lee. And the noose tightens. Maybe they could just install cameras and microphones on all newborns. Ludicrous, right? The world I grew up in back in the 50s and 60s stands in stark contrast to the amount of surveillance to which we are subjected now.
I suspect that remote activation of cell phone microphones is ultimately a better way for the government to record all conversations. However, I understand they like to have redundancy in their collection systems, just in case someone institutes an effective counter-measure against a particular technique.
One way of defeating spying is to stop communicating. But it requires some self discipline, so it’s not for everybody.
Anybody notice that page 21 of that promotional document contains the line:
“A continuous spool of up to 72 hours of audio data at each
sensor often contains court-admissible probative forensic
evidence”? That seems to directly contradict what the company is saying publicly.