快猫短视频

Voice jammer stops anyone from recording you speak

An artificial intelligence voice jammer can unobtrusively block microphones recording a single voice in an area, avoiding causing wider disruption that might tip people off
Artificial intelligence entity using voice to communicate
Using artificial intelligence to produce the right pattern of sound can render a voice unrecordable
Shutterstock/ArtemisDiana

A voice jammer can now stop anyone recording the speech of a single target person.

Voice jammers work much like noise-cancelling headphones, which effectively squash unwanted background sound waves out of existence by playing a copy of a background sound wave but with the wave pattern inverted. Such jammers generally stop electronic eavesdropping on conversations by broadcasting inverse sound waves that affect all microphones within earshot.

But this can prevent nearby people鈥檚 phones from picking up their own voices if they try to make a call, and can stop them interacting with smart assistants, tipping off everyone that something strange is going on.

Now at Michigan State University and his colleagues have created a voice jammer that uses artificial intelligence to isolate only certain voices before broadcasting the inverse wave, therefore causing no disruption to others nearby. They did this by training a neural network to isolate a certain speaker鈥檚 voice, and then used it to create an inverse sound matched specifically to their speech.

Their tool, called Neural Enhanced Cancellation (NEC), goes one step further, though. Rather than transmitting an audible sound, which could affect other people, it makes use of a bug found in all but the most expensive microphones. This bug introduces sounds at set distances above and below the microphone鈥檚 recording frequencies. NEC is able to use this quirk to play its inverse speech in the ultrasonic range, so that humans can鈥檛 hear it. Doing this at the point of recording provides the necessary frequencies to secretly block an audible voice.

When NEC is used in a busy setting, there is no humanly audible sound and no indication for anyone nearby that jamming is taking place, but it effectively makes it impossible for any microphone to record the target. The team tested NEC on a range of smartphones from Apple, Xiaomi and Samsung at a distance of up to 3.6 metres and found that it effectively blocked voices.

Yan says creating an unobtrusive jammer is vital to avoid drawing attention to it, and to prevent disruption to nearby people鈥檚 calls. 鈥淭he most common thinking is 鈥榡ust jam everything鈥,鈥 he says. 鈥淟ike, take a microphone jammer and jam everyone. But this is too intrusive.鈥

鈥淚 think for some people, like government officials, if they want to protect their voice from being recorded, this type of technology is definitely very useful,鈥 he says. 鈥淲hen you want to have a private conversation in a public environment, but you want to make sure your conversation is private and not eavesdropped by a hidden microphone underneath the table.鈥

Reference:

What gives humans the advantage over our incoming robot masters?
Topics: AI / security / sound