The Cub Reporter

The Vulnerabilities of Voice Recognition

Siri+Voice+Recognition
Back to Article
Back to Article

The Vulnerabilities of Voice Recognition

Siri Voice Recognition

Siri Voice Recognition

via Flickr

Siri Voice Recognition

via Flickr

via Flickr

Siri Voice Recognition

Hang on for a minute...we're trying to find some more stories you might like.


Email This Story






A surprising amount of electronics today use voice recognition software. Companies such as Amazon and Google have even made devices that can control a house with the power of voice recognition. But how does it work? Voice recognition software may be simple to use, but it’s not as simple on the inside. When a person speaks, they make vibrations that travel through the air. The device then picks these up and converts them into a digital string using an analog-to-digital converter, or an ADC for short. The system then tries to determine what sounds are voices, and which sounds are background noises. It filters out the other noises, and keeps the voices. It then adjusts the speed and volume of the sound so that the computer can understand it. The next step is to divide the speech into different segments, each being only 100ths of a second long. It also determines what sounds are plosive consonants. Plosive consonants are sounds that linger, such as “puh” and “tuh”. The program then sorts all of these sounds together again into a recognisable string. It then searches through its database to find a match for these words.

While the technology behind voice recognition is incredible, there are several flaws. For example, people use Apple’s Siri and Android’s Ok Google every day. But what a lot of people notice is that is has no clue what they said. This is because its software failed somewhere while translating the voice. Bad translation isn’t the only problem with voice recognition though.

Thousands of people every day record their voices on websites and apps such as Youtube and Facebook. It is easy for hackers and other people with malicious or devious intents to record these voices and use them. However, it is far easier to record someone’s voice with the help of voice recognition software. A video on Youtube can be taken down, and so can Facebook posts. However, it is far harder to “take down” a voice from a device with a voice command feature. These devices keep a permanent record of everything said to the device and how it was said their archives, so they can recognize who it is. This is handy for when a device recognizes a voice so that it can turn on, or so that only the owner can use it. But when a hacker goes into the device and gets it’s voice archives, they now have an entire list of every word  ever said to the device or around the device while it was on.

Many people have never of heard digital voice impersonation, but it is a true problem. Thousands of people every year are impersonated by someone who has access to their voice. They use other people’s voices to commit crimes and ruin reputations. They can also use them to get into people’s personal files and convince other people online that they are someone else. This form of identity theft is just as common as other forms, if not more common.

Voice recognition software may have many dangers, but that doesn’t mean that one should never post a video or use voice command software. Just be careful when using voice command.

Print Friendly, PDF & Email
About the Writer
Colin M., Co-Editor in Chief

Colin M. is an 8th-grader at Kraemer Middle School. He has an older brother and sister as well as a dog named Chloe who he loves very much. His brother...

Leave a Comment

If you want a picture to show with your comment, go get a gravatar.




*

Navigate Left
  • The Vulnerabilities of Voice Recognition

    Science & Technology

    DELTA RUNE- a Sequel, Or Something More?

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    Two Satellite Galaxies Conflicted ‘Recently’

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    Google Pixel 3

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    Science behind the Lottery

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    Researchers Discover Two New Chinese Dinosaurs: Bannykus and Xiyunykus

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    Elon Musk’s Hyperloop

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    How important is sleep?

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    Tesla Model 3

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    Phobias and the Science Behind Them

  • The Vulnerabilities of Voice Recognition

    Science & Technology

    The Truth about Bees

Navigate Right
The student news site of Kraemer Middle School
The Vulnerabilities of Voice Recognition