Last May we made a joke about Google creating Skynet in our blog post.
It turns out we might not have been that far off.
Recently Google announced that it is applying for six patents that focus on the fundamental building blocks of Artificial Intelligence and machine learning. One of the most notable is a technique developed by Alexander Krizhevsky, Ilya Sutskever, and Nitish Srivastva of Toronto University that focuses on training neural networks. This is technique that is used very frequently by AI researchers.
Other patents include image recognition and classification as well as reinforcement learning for computing systems. Both of which are geared towards learning and development and already tie in well with Google’s neural network whose psychedelic images you may have already seen.
Some are speculating that Google is doing this to hedge out competition in an effort to secure their spot at the top. Others seem to think that this move will significantly hamper research into AI technology, slowing it down or perhaps losing control.
Any way you look at it, Google appears to have taken the most significant steps to date in patenting AI technology. This can lead to the competition doubling the efforts, effectively starting the AI race.