There are multiple conversations from shot spotter admitted into court. New Bedford, Connecticut and Oakland had court cases off the top of my head, but a much wider range of conversation ends up being listened to by police and parallel construction is a thing.

Further, the concern is unequal policing biases statistics resulting in more uneven policing. A grid of microphones covering a full city is unbiased, placement based on past data isn't.

Drug crime is known to be fairly evenly distributed across different incomes, arrests are extremely biased.

No, in New Bedford, the "conversation" was not admitted into court [1]. Oakland is the "one" case I mentioned [2]. Do you have a source for the Connecticut case?

[1] "New Bedford utilized ShotSpotter, and the following was recorded by the gunfire detection system: “‘Oh my God! You're crazy!’ and then ‘Jason don‘t!’ several times, followed by a number of gunshots. After the gunshots, a female was heard yelling ‘You . . . missed and they shot him!’ ‘You're going to jail!”’. A motion to suppress this evidence was granted in Denison’s case, as the judge ruled that the ShotSpotter recording violated the Massachusetts Wiretap Act" https://readingroom.law.gsu.edu/cgi/viewcontent.cgi?article=...

[2] https://casetext.com/case/people-v-johnson-5116

That's what I get from faulty memory. However, evidence is still being used even when it's not directly brought up in court.

"In the 2015 Massachusetts case, the court refused to allow the introduction of ShotSpotter audio after the defendant argued the audio was captured and obtained in violation of that state’s wiretap laws. Police in that case requested and received an extended audio clip from the company that captured people discussing the shooting, including one calling out the defendant’s name."

In theory any evidence from those audio recordings is tainted and no longer admissible, therefore police should want to avoid listing to any of it. But... by not bringing it up in court police can use parallel construction to get around those issues.

Those "extended audio clip[s]" don't seem to be available anymore as a matter of policy after 2019, following an external privacy audit done by the Policing Project at NYU Law. It also seems that police don't have access to audio outside of gunshot incidents.

https://www.policingproject.org/shotspotter

https://www.shotspotter.com/wp-content/uploads/2021/01/ShotS...

edit: Good point, I agree that parallel construction is an important concern, especially since there's so little transparency in police investigations. One possible solution is to force them to limit the duration recorded to only a few seconds. (I previously thought that's how ShotSpotter worked. Turns out they've kept days of audio before, apparently for "debugging purposes.")

> Drug crime is known to be fairly evenly distributed across different incomes, arrests are extremely biased.

Keep in mind that the reverse is true for gun crime, where low-income minority communities suffer from the lowest clearance rates.

Audio evidence is being used such as in the Massachusetts case.

PS: You are quoting my previous post not an edit: "Drug crime is known to be fairly evenly distributed across different incomes, arrests are extremely biased." is from further up the thread.

Yes, I know that part is a previous post. You also heavily edited GP, but current state of comments is reasonable now.

Isn’t this whole article about shooting though and not drug crime? How would you hear drug crime?

Are shooting crimes known to be fairly evenly distributed across different incomes?

ShotSpotter is recording conversations inside peoples homes and on the street. You can definitionally hear someone making a drug deal or talking about the joint their smoking etc.

I'm reasonably concerned that once recordings exist getting access to them is much easier, especially as police have already heard many conversations taken from these recordings.

Can you point to an example of a ShotSpotter sensor recording a conversation inside somebody's home? That seems unlikely given that these sensors are installed outdoors.

They are installed outdoors, sometimes on top of people's homes.

https://twitter.com/loteck/status/1760859493133697237?t=AQTP...

Are these things public microphones that pick up street conversations or are they specialized microphones without stored recordings? The capability of these things is the only important question, since if they can be used for mass surveillance we shouldn’t trust cops with them.

My guess would be that since they need to record a shot without knowing when it happens, and then forward the sound of the shot to an expert to confirm it was gunfire, they must have the capability to locally listen and record continuously, and we can hope that they only keep a few seconds of recording before/after the event.

that doesn't really change the fact that they have microphones that are capable of recording voices, they have storage of the recordings, and are able to retrieve recordings and are willing to provide them to the police.

Yes it is very important. Shotspotter is like a DVR. It's always "recording", but it is running on a very very short recording loop.

The detection of gunshots is based on an algorithm. When triggered, it saves a few seconds before the sound and several seconds afterward.

It is different than police just being able to pull up the recordings for a certain street corner from 2 weeks ago or even a day ago and listen in on people's conversations. The case in question came from a shotspotter recording gunshots, and the short clip also happened to include speech within it since it directly followed the sound of shots.

ShotSpotter's own public statements say that audio is retained for 30 hours, see e.g. https://www.soundthinking.com/faqs/shotspotter-privacy-faqs/. In the past they provided Albuquerque with a document that said 72 hours, but I believe they have been reducing that period over time. It may also depend on their specific contract with the law enforcement agency. Still, we know that they must retain recordings, because they do take requests from customers to reanalyze data that did not result in an alert.

For convenience, the most relevant text: "If the system misses a gunfire incident, police may contact the company to see if there is any audio or location evidence. In this case, only authorized ShotSpotter personnel with proper credentials can access sensor audio to search. Their search is limited to the 30- hour sensor storage timeframe."

ShotSpotter used to have significantly fewer privacy protections, and retention was indefinite early in the product's life. Fortunately, a combination of legal challenges and statutory privacy policies among their customers have lead to them significantly improving their privacy controls over time.

Is that, like provable though, thats my biggest issue. Cities are allowing a private company to mass deploy microphones all over (mostly lower income) areas and just _trusting_ them to not be keeping recordings.

> it is running on a very very short recording loop

I'm curious what your source for that information is

They made it up

How would we know if this is true? For instance, is the source code subject to independent review, with verifiable builds made from that source so that all deployments have to show matching hashes? Or is more of a pinky-swear kind of a thing? If so, is the pinky swear even documented somewhere?

I mean, yah, it makes sense the one they admitted was pertinent to the case? I don’t see how that context is meaningful.

I actually strongly disagree with you — the context doesn’t matter. We have a private quasi-law enforcement entity installing thousands of surveillance devices in American cities without any external oversight or knowledge of where they are installed. These surveillance devices that were pitched as tools to locate gun crimes all of a sudden record audio? And this quasi-law enforcement company with no oversight is storing that data and then furnishing it to the police?

We have no idea what’s recorded, we have no idea where these devices are, we have no idea who is listening to the recordings, we have no idea what access LEOs have to these recordings, we don’t know how they are stored, and we don’t know how long they are stored for. You’re seriously okay with a non-government entity operating like this?