Yes, that's right, Internet giant Google's artificial intelligence division DeepMind has created a version of AI with its very own algorithm which protects messages shared between robots Alice, Eve… and Bob.
The @google|s #Brain init|s #AI Eve, Bob and Alice just created their own sec #algorithms|Did someone say #Skynet? https://t.co/pOPilCxHUs pic.twitter.com/G2N66tVXXV
— Max Mayerhofer (@Weyrother) November 1, 2016
This means Google has cracked the code enabling two artificially intelligent robots to communicate with each other without prying ears being able to listen in and disseminate their algorithms.
The bots didn't have much of a head's up either. They weren't pre-programed with information on how to encrypt stuff, they were set to function failure and left to their own devices.
The networks, Alice, Bob and Eve, subsequently created a generative adversarial network (GAN) between them. GAN basically means unsupervised machine learning, which is when it all gets a bit complex.
Basically, Alice and Bob were eventually able to create encryption tricks that mean they could hide messages from Eve. This, according to a recent article in Big Think, is "a little worrisome."
"Alice and Bob are the first neural nets who've kept secrets from their programmers. That we know of," the article says.
On one hand, Alice, Bob and Eve are a remarkable development in the world of AI — on the other — the whole concept offers the possibility of a future in which robots will be able to keep secrets from their creators — humans — which could be used for more sinister purposes.
