Tinfoil hat time: imagine a massive database with recordings of every voice in the world. Each one is as unique as a fingerprint, one way or another. Now imagine software that’s able to search through all of those voices, compare it to another recording, and identify the speaker in a matter of seconds. You’ve just imagined VoiceGrid Nation, and there’s a pretty good chance that it’s already able to find you.
It’s developed by a Russian group called Speech Technology Center, or SpeechPro in non-Russian markets. It’s already being deployed in more than 70 countries (and countless state and federal agencies across the US, including the FBI and NSA), most openly in Mexico, where law enforcement is already collecting hundreds of thousands of ‘voice-prints,’ samples that can form a database to match recordings against. It only takes 5 seconds to search through 10,000, so it shouldn’t be hard to cover the whole country.
So, how does a government go about collecting a database of voices? That’s probably our biggest question surrounding this. Aside from voluntarily collecting a snippet of everyone’s pipes in action, the only thing we can think of involves doing illegal stuff. Obviously, the whole system would be pretty useless if you didn’t have a sizable chunk of the population (or at least a sizable chunk of the criminal population) to match it against, so that’s definitely one of the first questions that this whole thing is going to raise.
If you’re worried, you might find it comforting to know that there are lots of less nefarious uses for SpeechPro’s technology. It apparently works great for call centers, who can use it to identify and annoy the hell out of you even more efficiently than before. We can think of tons of other great applications, like… well, we’re not really sure.
But the good news is that you probably won’t have to worry about it, since you won’t even know about it. A lot like TrapWire, and the tons of other violations of our privacy that we don’t know about. There’s nothing quite so comforting as ignorance!
And just to be neutral… it’s possible that there’s nothing to worry about here. If there’s a transparent way to collect a sample of everyone’s voice that doesn’t involve wiretapping and recording our phone calls, maybe the world would be more open to such an idea. Kind of like a fingerprint database, only with voices. Or maybe we could just stick to using it to identify the voices of people who are already under suspicion. Unless that’s everyone, of course.
Source: Slate
Fingerprint Image by Shutterstock
18 Comments - Add comment