More

    Google’s Artificial Intelligence Creates it own way of Encrypting Data

    What happens when two persons what to talk to each other but there is a third person listening to your conversation, simple you talk using an language that you both understand but the third one doesn’t know or understand. But what if you’re using an computer anybody can convert the language and understand it there comes the encryption algorithms, When used only the person with the decryption key can understand it and the random person who is trying to snoop in cant understand anything. There are many ways and algorithms for encryption and to break this without an decryption key it would take ages.

    Researchers from the Google Brain deep learning project taught neural networks how to encrypt and decrypt messages though they din’t directly teach them the encryption algorithms that are existing. In a research paper published online researches created three neural networks named Alice, Bob and Eve and made them to pass messages by creating encryption methods themselves.

    Now, each neural networks were assigned a task: 

    • Allice had to send an message that only Bob could read

    • Bob had the job to decrypt the message by the given key

    • While Eve had to eavesdrop on their conversation and try and decrypt the message by herself

    As said before none of the neural networks involved in the study were taught cryptographic algorithms so they were not able to develop sophisticated encryption system though they were able to convert plain text into encrypted messages. At first Alice and Bob were bad at hiding their secrets. But trying them over 15000 times, Alice was able to create her own encryption strategy and Bob simultaneously figured out how to decrypt it. The message given was 16 bit long with each bit being a 0 or a 1. Eve was only able to guess 50% of it meaning that he was just flipping a coin with two sides so it was an complete guess work.

    Of course, the personification of these three neural networks oversimplifies things a little bit: Because of the way the machine learning works, even the researchers don’t know what kind of encryption method Alice made while encrypting her messages therefore it wont be useful in any practical applications.In the end, it’s an interesting exercise, but we don’t have to worry about the machines talking behind our backs just yet……


    We would love here from you about what you think about this in the comments section below. If you want to stay updated with our latest updates you can like our Facebook page or keep an eye on our website also follow us on Twitter @hacksandgeeks

    Latest articles

    Related articles

    Leave a reply

    Please enter your comment!
    Please enter your name here

    This site uses Akismet to reduce spam. Learn how your comment data is processed.