[kijin]

> the system was not intended to record anything except gunshots

So, they could record conversations, but they won't for the time being. A classic example of using software to temporarily cripple the true capabilities of a computing device.

In the hands of DRM-wielding corporations, "defective by design" results in inconvenience and loss of users' freedom. In the hands of the surveillance state, the same technique results in a situation where citizens must simply trust the authorities to exercise restraint. Because the police could flip a switch at any time and record all sorts of conversations. Somehow I don't trust that the switch will remain un-flipped for long. And when it does get flipped, as it did in New Bedford, everyone will say it was just an accident.

Maybe, just maybe, we should accept systems like this as a necessary evil in certain cities where there's a lot of gun violence. Still, I don't like this. When it comes to the government, I'd much rather give them hardware that can't be unlocked "by accident" because there's nothing to unlock. The thing is, it's unrealistic to do that in all cases, and we have to hit a balance somewhere. Which is exactly why systems like this raise difficult issues.

[tomerv]

Given that the microphones are placed in a public location, and that they record noises heard in the street, one could argue that it's a reasonable level of surveillance.

The real concern will be when governments start to identify their citizens by their voice signatures, and then use that to track their whereabouts.

[InclinedPlane]

That's not the worst of it. Consider the full implications here. Imagine a worst case scenario where gunshot detection is used as a pretext to put microphones all over city streets, every block say. Now imagine that those recordings are stored forever and properly time / location tagged, not a technologically difficult problem.

Then imagine what is possible with such data. First off, you could use triangulation and advanced filtering between multiple microphones to be able to pinpoint the source of each sound and separate it out from the background. You could, as you say, identify individuals by their voices. You could track their wherabouts. You could monitor who they are talking to and when. You could learn so much about their lives by monitoring all of their conversations in "public". In the worst case scenario of the government turning into a police state this is a frightening level of surveillance.

[tptacek]

Apart from the fact that the devices are apparently designed not to be capable of recording conversations, and can clearly be improved to make it even more difficult to record conversations, "not deploying gunshot detection" isn't the only privacy control that cities can employ; cities can just make it illegal to collect raw audio.

[kijin]

The article mentions that the devices did manage to record conversations in at least one case. If so, they are clearly capable of recording conversations, just not optimized for it.

Also, how do you distinguish gunshots from background noise and triangulate the location of shots without first collecting raw audio from multiple devices and analyzing it?

[bigiain]

Indeed, there's clearly some disinformation going on in the article:

"James G. Beldock, a vice president at ShotSpotter, said that the system was not intended to record anything except gunshots and that cases like New Bedford’s were extremely rare. “There are people who perceive that these sensors are triggered by conversations, but that is just patently not true,” he said. “They don’t turn on unless they hear a gunshot.” "

So apparently "the sensors", "They don’t turn on unless they hear a gunshot.". How, exactly, do they "hear a gunshot" if they're not (yet) turned on?

I suspect the truth is there's some software configuration that inhibits _recording_ of the sensor data until a gunshot-like event occurs (though if _I_ were designing this system there'd be at least a 30second or so buffer, so I could archive the sounds that if heard _before_ a gunshot as well as afterwards). But I'd hesitate trust that "configuration" to be particularly secure - much like the TSA "pornoscanners" - which in spite of claims of it being impossible, seem to be able to record images for the amusement of the operators and their friends…

[abecedarius]

They could be designed to run some local preprocessing and only record/forward audio when a gunshot seems significantly probable. (Not that I'd bet they actually do. We need public pressure to get them made that way.)

[underwater]

If I recall correctly, privacy laws are based more on reasonable expectation of privacy rather than where the recording device is located. I can't climb a telegraph pole and photograph someone inside through a second floor window, for example.

I don't know how this would affect the legality of surveillance though.

[Jach]

>The real concern will be when governments start to identify their citizens by their voice signatures, and then use that to track their whereabouts.

And thanks to a recent heavy-hitter at the box office, the public has already had it decided for them that it's unethical! (Unless it's used once for a really really bad guy.) But apart from that this is an awesome system for its purpose. Keep it out of my home (they've already got phone-tapping for that anyway) and I'm fine with it.

[wl]

ShotSpotter cannot be "unlocked" to record conversations by accident. These sensors are placed far away from where people might be talking and this is intentional. Conversations are just one more source of noise to be filtered out.

[kijin]

Engine sound is another source of noise, but ShotSpotter can record engine sound and use it to figure out whether the shooter was in a car. Why would conversations be any different? Noise filters can be turned on and off.

This is no different from modifying the nude scanner at the airport to display a generic human form rather than the actual outline of your body. The machine still has the capacity to display what your private parts look like.

[vl]

Most likely they use the fact that the shot location moved, not the engine noise.

[nosse]

How? With multiple shots it's easy as hell. But from one it's quite hard?

You could hear the firing pin hitting the cartridge (harder to hear than engine running) then you should hear the gasses exiting barrel somewhere around 0,001 second later. Then calculate doppler... Sounds kinda hard.

[bigiain]

I bet the doppler shift from two different listening microphones would give you a pretty easily calculation for how fast the gun was moving (at least relative towards/away speeds with respect to the two microphones) - I'd guess 3 or more microphones listening to the same gunshot could pinpoint both direction and speed.

[nosse]

I was thinking about the Doppler too, but the Doppler shift of the bullet while in flight doesn't really do it. And the time-frame to do it while inside gun is really really small.

[redthrowaway]

I highly doubt they're providing that service for single shots. Besides, I can't see too many thugs firing a single shot from a car in commission of an assault or homicide. They're not exactly known for their accuracy at the best of times.

[kijin]

Ah, that makes sense.

[Tsagadai]

Locks are mostly superficial. I've seen too much footage recorded from security camera monitors with phones to think that any monitoring system is secure or locked when people are involved.

On the growth of such systems: it is almost inevitable that sensors end up getting placed everywhere. You have hundreds of sensors in your phone right now, many of them are uploading data to websites at various times.

Sensors will be placed on almost every utility as a matter of safety and efficiency. It is important to know when a pipe breaks or if there is a small leak in a gas main. Look up any given utility company and they will probably have some "smart" this or "integrated" that scheme in development.

There are cameras everywhere in most cities now, it is a matter of time before there are microphones too. Identifying who is where and who is speaking is a programming problem that has mostly already been solved. The two will be put together in the interests of safety and fighting crime. Someone might even invoke the terrorism word to speed up the process. Small-scale at first, in key places and important settings. Then rolled out across entire cities and states.

I am almost entirely visible to security cameras from the moment I step outside in the morning until I reach my office. Eventually, these systems will be integrated and process data effectively. The question is, what happens then?