Sri Deivasigamani - Would You Let an App Listen for Gunshots?

SplitSec.AI founder and CEO Sri Deivasigamani joins Lars to discuss technology that turns smartphones into sensors capable of detecting nearby gunfire. Would you give an app access to your phone’s microphone if it could help police respond faster? 

Transcript

Copy Transcript Copied

SPEAKER_00 0:19

Hey, welcome back to the Lawrence Show. So a Chicago startup company wants to use your smartphone to become a gunshot detector, tracking shooters in real time. Should your phone be listening for gunshots 24-7? I thought we'd talk about that with the founder and CEO of a Chicago-based company called splitsec.ai. And that's Sri Saga Deva Saga Sagomini. And uh I have a I hope I haven't murdered your name too badly, Sri. Welcome to the program.

SPEAKER_01 0:50

Yeah, I'm excited to be here. Thank you for having me.

SPEAKER_00 0:53

So tell me about this app that you want people to load on their uh smartphones.

SPEAKER_01 0:59

So what this is is you know it turns ordinary smartphones uh into AI-powered gunshot detectors. And the model runs on the phone, entirely on the phone. All the audio analysis is done on the phone, totally private. Nothing ever goes to the cloud. And if it detects possible gunfire, it can warn the user and nearby people within seconds.

SPEAKER_00 1:20

And it does this sort of passively, in that most of our phones, I mean, whether we like it or not, are listening to us all the time, unless we have them physically turned off. And and they may even be listening then. What can you tell us about that?

SPEAKER_01 1:35

Yeah, we we are a very privacy-centered company. So the audio is analyzed on the phone itself. No audio ever leaves your phone. It's a couple second audio or three to four seconds, and it's just discarded. And all that analysis is done, and we have designed it specifically that way for privacy reasons.

SPEAKER_00 1:56

And if you have this smart app loaded on your phone, and presumably other people in, say, a place like Chicago also have it uh loaded on their phones, uh, that at what point does the detection of gunshots by your phone become something that is shared with the folks who have the app loaded on their phones, say, not too far away from where you are? Is that the basic idea that if if a number of phones, all in a given area, all detect gunshots, compare notes and say, yep, this looks like gunshots, and then the people with that app loaded on their phone get a warning?

SPEAKER_01 2:33

Yes. I mean, uh the this is where the you know having a network of phones becomes powerful. Uh we need really w one or two phones to detect the gunfire. And we are able to then distribute that alert within seconds to everybody who is within one mile radius.

SPEAKER_00 2:52

I'm talking to Sri Deva Sagamini, who is founder and CEO of Chicago-based splitsec.ai. How do you plan to roll this thing out? And what are you going to consider successful uh if you get people to adopt this? I mean, do you have to have a great many people in any given city adopted to be successful?

SPEAKER_01 3:12

Uh actually, uh no. You know, if you think about Chicago, uh, you know, if you think about different wards, each ward has a population of about 50,000. About 1 to 2% adoption uh gives us fairly large coverage within that ward. And so that that is that is the type of coverage uh you know we are looking for. And again, you know, we are doing pilots right now and we're taking feedback, we are trying to figure out what works and what doesn't. Uh personally, success uh for us would really mean we have warned somebody and somebody's gone to safety. Uh, that is a huge win for the company. Uh, you know, we have had users already confirmed, they listen, they got actual gunshots and they've confirmed it.

SPEAKER_00 3:53

I I'm curious about this though. How much does my phone hear if I've uh say I say I'm outside, but my phone's in my pocket, which it's often in my pocket, front pocket or back pocket, whichever it happens to be. How much can that microphone, that tiny little microphone, on a phone that's inside my pocket, how much can it hear?

SPEAKER_01 4:14

So I can speak specifically to gunfire. Uh, you know, gunfires are uh high energy, uh, you know, they have a sharp peak and they fall. And so that is actually carried quite far and to a lot of things, including, you know, things like trees and buildings do obstruct and reflect it, but they are carried quite far. So even if you have the phone in your pocket, and let's say there is a gunfire incident, uh, you know, at the at the next traffic light, uh, it's very likely that your phone would would uh pick that up.

SPEAKER_00 4:45

Would it also give people the ability to know which direction to run to run away from the gunfire?

SPEAKER_01 4:51

Exactly. That is exactly what you're doing. So uh, you know, when we have uh, you know, when two phones pick up gunfire, two or more phones, we have the ability to uh roughly see where the gunfire came from. Uh and so we'll be able to tell people look, we can't give you the precise down to the feet, but we can say it's it's coming from this area, so you got to go in the other direction so people don't run into gunfire.

SPEAKER_00 5:18

Uh Dave, uh uh Sri, I I've talked to people from the company called ShotSpotter before, which I think was it deployed in Chicago as well. And the police who've said it it's actually very, very useful to them, but an awful lot of bigger cities uh will follow the politics and they've gotten rid of ShotSpotter because they said somehow it discriminates against certain ethnic minorities and that sort of thing. But at some point does SplitSec and its app end up sharing this information with the police to make sure the police can also respond to the sound of gunfire?

SPEAKER_01 5:55

So we uh you know we are not sharing anything with police. Okay. We are solving a different problem. You know, traditional systems, they primarily notify authorities. What we are designed to do is give ordinary people uh you know nearby immediate situational awareness. That is our goal.

SPEAKER_00 6:14

And nothing beyond that at this point.

SPEAKER_01 6:19

Nothing uh you know, we don't have any integration with uh with the police. Uh our technology is modeled similar to WAVE in the sense that you know you you give and take for your own community. People open up WAV to see if there's a traffic jam up ahead, uh, or if there is emergency vehicles going ahead. Uh so our view is this app, uh, you know, you open this app to see if anything's happening nearby. And if your phone picks up something, you make a contribution. Uh now, presumably the police uh could download the same app and look at uh you know what is happening in certain neighborhoods or you know, wherever shots are being reported. But whatever the information they'll see is totally anonymous, reported by people, nobody's uh location or address is revealed. Uh so that much uh the police can do today.

SPEAKER_00 7:06

Well, it sounds it sounds like it's gonna be interesting. How soon do you expect to have some concrete results from this experiment in the uh 21st and 42nd wards? Ah, we are learning every day.

SPEAKER_01 7:19

So the app is out. Uh in fact, you know, we noticed that even people outside of Chicago have started to download the app. Uh it's available on Android. Uh, we are learning many, many things. You know, we are learning, for example, on some phones, uh, you know, the battery life has gone faster than we expect. Uh, some phones have produced more false alerts than we like. Their microphones seem to be sensitive. Uh, so we are learning all these things and then implementing changes uh very, very rapidly.

SPEAKER_00 7:46

Well, I appreciate you coming on the program. That's three Davis Sagomini, who is uh the founder and CEO of a Chicago-based startup called splitsec.ai. And we appreciate him coming on the program. By the way, take a listen. Did you hear that? Listen very carefully. That's my phone. No spam calls, no spam texts. I love that sound of silence. Your phone, maybe not so quiet, but you can make it more quiet. And here's why the data brokers are selling your personal information. And they used to do that with my information until I employed incognite. Why? Because the data brokers gather your phone number and your home address and your email and family members and political affiliations. Even where you do your banking and those scammers blowing up your phone, they didn't get lucky. They bought your information, and they're going to keep on buying it over and over until you make them stop. I did that starting in January of this year with Incogni. INCO GNI. Incogni automatically removes my personal info from hundreds of data brokers and people search sites, then sends the legal removal request. If they can't find you, they can't target you. Get 60% off now at incogni.com slash Lars. That's INCO GNI, incogni.com slash Lars.