Take for instance IBM’s Watson, it is a type of AI, but all it does it make surmises and give you answers based on some sort of statistical understanding and machine learning.
We’re still at the dawn of the AI-Age. We are yet to define exactly what ‘intelligence’ is for humans. How can we define exactly what ‘artificial intelligence,’ is?
When it comes to computer programs, they can be taught to learn for a very specific purpose using Machine Learning. Often, people confuse machine learning with human learning. Now, machine learning is simply training the software to behave in a certain manner based on a statistical pattern. Everyday objects such as self-driving cars, Watson and even image recognition software, are taught to behave in a certain manner and cannot go beyond those limitations. They don’t really pose a threat to humanity.
But, is it possible that in a few hundred years software become sentient? It’s possible, but machine won’t be intelligent enough to pose a threat to humanity. You see, a machine would never have the same emotions that we do – anger, greed, envy, etc. Even if a piece of software is highly intelligent, there’s no reason why it would necessarily be a threat to humanity. If anything, the software would be controlled by us. All technologies have both good and bad aspects that can be used for. Take for instance nuclear, which is the most glaring example of this with nuclear bombs and nuclear power
The reason why researchers and scientists consider AI to be potentially dangerous is because intelligence could increase to colossal levels. Moreover, it ca