I myself have heard the words, “Artificial Intelligence is going to destroy humanity” many times. Whether I am at my job, with friends, or online, it seems that many people are under the wrong impression that somehow the new advent of robotics and the integration to concepts such as Computer Vision, Machine Learning, and Artificial Intelligence will follow an exponential curve resulting in the singularity, or the point in time when Artificial Intelligence meets the intelligence of an average human.
Now, I’m not going to say that this will never happen. It definitely has a chance of happening in the future, but definitely not the near future. Let’s take a look why.
Right now, Artificial Intelligence is primarily geared towards Convolutional Neural Networks, Generative Adversarial Networks (GANs), and Deep Learning. We classify all of these different methods of machine learning to be artificially intelligent, which they are, but only to an extremely limited degree. Whenever we code a GAN or a Neural Network, we are programming the system to learn over time, commonly segmented into what is known as Epochs. However, the system is not learning by itself! We are forcing it to learn.
One important facet to what distinguishes us as human beings is our emotion, or personal connection to ideas thrown at us. If you were a kid and your mother told you to take out the garbage, you would not always take out the garbage. Maybe, you would say, give me 10 minutes, or start whining. What we are doing with robotics is automatically forcing the robot to learn, then automatically evaluating the machine after.
Now, some will argue that this is the reason that the field is known as artificial intelligence, because it is artificial, not generally intelligent. That is absolutely a very strong point to bring up, but lets compare what the word artificial really means. Taking an analogy, we can look at Oranges and Orange Candy. Now, an Orange is not artificially flavored, but Orange Candy is. What makes Orange Candy artificial is that it is not occuring naturally. Nevertheless, Orange Candy still tastes just like an Orange, at least most of the time.
We can use this analogy to draw a comparison to how we see Artificial Intelligence. If we wish to define a field such as Artificial Intelligence, are we not to carry out the same processes that regular intelligence does, but simply not let it occur naturally? That is why in Artificial Intelligence must be re-defined.
Overall, it is easy to say that the true definition of Artificial Intelligence does not matter, but thinking that is not gripping the problem by the horns, and dealing with it. Until America sees that the current definition of Artificial Intelligence can never reach the singularity, our economy will not be able to move as fast as we are potent to. If we choose to continue to grow as much as an economy as we are currently, we need to capitalize on true AI Applications.
If you enjoyed this article, please follow this Medium Channel to learn more about Tensorflow/Machine Learning
Source: Deep Learning on Medium