Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

ShotSpotter requests to be held in contempt rather than disclose its methodology

chicagoreader.com

Post reply on HN
Print view View original

191–200 of 320 posts

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#191

Post by pseudo0 »

flaviut wrote: post #15
pseudo0 wrote: post #8

Earlier quoted context omitted.

Would results of discovery be public record? If so, I could also see: 3) They don't want their "secret sauce" made publicly available. That would both open them up to commercial competition, and make their product less reliable in court, as anti-Shotspotter experts would start figuring out how to effectively argue against the product's methodology. In general I think technology like this should have to be fully trans…

This is a reasonable argument, but I'm not sure you're aware of allegations of this company altering evidence to better fit a police narrative: https://www.vice.com/en/article/qj8xbq/police-are-telling-sh... I sincerely doubt there's any secret sauce to protect when they have "analysts" sitting around regularly "correcting" evidence.

I agree that there is definitely moral hazard when it comes to after the fact "corrections", particularly if the analyst knows the outcome desired by police. The company's track record looks pretty poor in this respect.

In general though, I think there are legitimate reasons for manual analysis. Any machine learning approach to a problem like this is going to have to balance false positives and false negatives, and there is necessarily going to be a somewhat arbitrary cutoff. Detecting bang-like noises in a large city is likely going to have a pretty conservative cutoff to avoid DDOSing the police with calls. For example say the company sets their cutoff for automatic reporting at 90% confidence, but when the police ask them to review a specific time period, it turns out there was a shot detected with 89% confidence, and manual analysis indicates a false negative. That is probably still useful information, and it would definitely be useful to include this data point with the correct classification in future iterations of the model.

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#192

Post by woodruffw »

remflight wrote: post #183

If they don’t reveal their methodology then how could this be used as evidence? Shouldn’t that lead to the entire case being thrown out if the police were alerted by this? Seems ridiculous to not allow the defense to see the methodology, even radar guns get that treatment.

> If they don’t reveal their methodology then how could this be used as evidence?

It isn't used as evidence; it's used to manufacture pretext to justify collecting "real" evidence. SpotShotter is not an ignorant party: they understand that the service they sell to police departments is a pretextual laundromat for busting up whoever the police feel did the crime.

That's the crux of this case: SpotShotter is being asked to produce evidence that they are compelled to produce, and they'd rather take the loss in this instance (and continue to use it for parallel construction) elsewhere.

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#193

Post by woodruffw »

nickysielicki wrote: post #55
solardev wrote: post #36

Earlier quoted context omitted.

I dunno if this company deserves that benefit of the doubt. Honestly, given their history, they probably just don't want to reveal that their entire business is snake oil, basically existing solely to make money by enabling law enforcement to create crimes out of thin air. If their science was sound they wouldn't have to resort to such tactics...

Why are you so skeptical of their technology? Look how accurate raspberry pi powered ADSB plane tracking networks are — and that’s a much harder problem: they have no funding, they’re doing triangulation at the speed of light, and they’re doing it without GPS clock sources. You could also point to the lightning tracker networks that work in roughly the same way. Given a few million dollars to outfit their nodes with…

Just to point out: on a signals processing level, I suspect that monitoring a signal that's designed to be tracked is much simpler than triangulating a sound using an incomplete acoustic profile. We can track planes via ADS-B with consumer hardware because it's a constrained and well-defined problem that was designed with automation in mind; this is not the case for the noises made by guns in an unknown environment.

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#194

Post by calvinmorrison »

public_defender wrote: post #129
calvinmorrison wrote: post #121

Earlier quoted context omitted.

But ShotSpotter doesn't do stop and searches. If police are using this to justify searches, thats between the courts and police, not shotspotter

The issue is that ShotSpotter may be profiting by giving police an excuse to search people when they would not otherwise have sufficient legal grounds.

Again sounds like the issue is between the police and the courts over what constitutes sufficient legal grounds.

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#195

Post by shmde »

Soo, can lets say someone work with the court, shoot random shots and fireworks around a city. Give the court their exact locations from which shots(gun and fireworks) were fired and ask ShotSpotter to identify gun sounds to that of fireworks and also figure out the locations.

If ShotSpotter is able to figure it out correctly only then they are are allowed to NOT disclose their tech otherwise it would be very illegal to keep it undisclosed because people can literally get arrested on the wrong leads generated by a piece of tech.

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#196

Post by esoterica »

quickthrower2 wrote: post #177
esoterica wrote: post #156

Earlier quoted context omitted.

The magic box already exists, it's called people who call the cops to report gunshots every time they hear fireworks are a car backfiring. If you think it's wrong for police to investigate a shotspotter alert because it might be a false positive then the logical conclusion of that line of thinking is that it should be illegal for the cops to show up to investigate when someone calls in to report gunshots, since those…

I don’t see the logical connection between what the parent comment said and your reply. He is saying they could be selling pure snake oil, and worse using it for discrimination (but not sure what the basis was for that, if any).

Shotspotter stores the raw audio so you can just ask for the audio and evaluate for yourself whether a reasonable person would identify that as a gunshot.

Either hearing a gunshot is not probable cause to pull someone over (in which case Shotspotter's accuracy and methodology is irrelevant), or hearing a noise that sounds sufficiently like a gunshot is indeed probable cause, in which case you can just listen to the audio yourself and make a determination about whether or not there was a true gunshot-like noise or the company invented a fake alert out of thin air. I don't see any scenario in which the methodology or accuracy of Shotspotter's devices is at all relevant to this case. The relevant fact here is whether this particular bit of audio can be reasonably classified as a gunshot, not the broader accuracy of the system over the whole data set.

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#197

Post by YLE118 »

ta988 wrote: post #171
YLE118 wrote: post #136

Earlier quoted context omitted.

An egregious example from the first link: Police said ShotSpotter, a surveillance system that uses hidden microphone sensors to detect the sound and location of gunshots, generated an alert for that time and place. Except that’s not entirely true, according to recent court filings. That night, 19 ShotSpotter sensors detected a percussive sound at 11:46 p.m. and determined the location to be 5700 South Lake Shore Driv…

The microphones are not hidden you can see them, they are just discreet.

Kind of like speeding and red light cameras, though they could be placing small enough microphones that you wouldn't see them while driving/walking by.

Of course that seems like a small nitpick given the content of the passage I copied from the article.

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#198

Post by xenadu02 »

nickysielicki wrote: post #130
alasdair_ wrote: post #69

Earlier quoted context omitted.

Imagine I made a magic "crime detection" box that called the police each time a crime was "detected". Now imagine all it did was randomly pick a number from 1 to 100 each day and if it were a 1 or a 2, it "detected" a crime, giving probable cause to police to search people near the box. Now imagine I almost exclusively put the boxes in areas where black people without a lot of money lived. You could see this would be…

I would have a problem with that, but I don’t think that’s what’s going on here. My gut feeling is that shotspotter mostly works and is mostly accurate, regardless of what Vice reported. I can imagine how I’d build such a system and I don’t think it would be very complicated. You don’t need any powerful signal processing or AI, this entire idea was feasible 20 years ago. And secondly, I think that if any technology i…

The problem is not that no one could build a system to detect and triangulate gunshot sounds.

The problem is that this specific company seems to be strong on marketing claims and light on engineering legwork. There have also been some implications that point to (in my opinion) post-indictment editing of records to gin up post-hoc probable cause in some cases. No smoking gun (ugh) to prove it but concerning.

What is known is that ShotSpotter reports seem to rarely create useful leads for police. It would hardly be the first (or second, third, fourth, et al) time someone promised to revolutionize policing only to hand over snake oil. Then when the money faucet begins to dry up some companies just can't help fabricating proof of their technology or even falsifying evidence to make sure police "see the value" in continuing to pay for it. Most police departments are being taken for a ride. Some fully understand the wink-wink-nod-nod going on and appreciate the company's "flexibility".

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#199

Post by brewdad »

trs8080 wrote: post #56
nickysielicki wrote: post #33

I don’t understand the argument here. Suppose shotspotter is deeply flawed and often confuses fireworks and car backfires for gunshots, how could that have any impact on this case or the legality of it? I don’t see it as any different than if a police officer is out patrolling, hears a car backfire, and confuses it for a gunshot. If he then detains someone as a part of his investigation and discovers some other crime…

If the officer lied about having probable cause (e.g., by telling ShotSpotter to fabricate alerts or to re-classify an alert after a stop to manufacture probable cause), the case can be thrown out.

But we can't know if the officer lied because SpotShotter refuses to reveal how their black box works. Since the officer's word is like unto God's in most courtrooms, the defendant is kinda stuck.

Re: ShotSpotter requests to be held in contempt rather than disclose its methodology

#200

Post by pxmpxm »

bergenty wrote: post #159
pxmpxm wrote: post #126

Earlier quoted context omitted.

Quoted post unavailable.

It’s not really that simple though. You can always go back less than 40 years and point out that street gangs were a response to police harassment. That communities were selectively devastated with targeted infusions of dangerous and addictive drugs. That there is a continuous pattern of profiling based on superficial attributes. A bad summer can ruin a person’s life, I don’t understand how we don’t expect generation…

The explanation I had in mind was more directed to the narrative over reality when it comes to shotspotter, rather than the mechanics of structural poverty incentives that invariably drive people toward illicit income that's fraught with violence.

We've had 50+ years of various great society type programs that are designed with negative net-of-transfers earned income gradient, and those were squarely aimed at racial strata, but it's a huge stretch to claim that a thing that listens for gunshots in a place where tens of people get shot weekly is some sort of bigot master plot.

Also tangentially speaking, anyone seeing a race dimension in "subsaharan africa life expectancy" is merely looking in a mirror.

Post reply on HN