Are robots out to get us?
Artificial intelligence (AI) has been around for quite a while now. At least in technology years where everything gets old pretty quickly. Classical philosophers tried to explain for the human brain works, in terms of thinking. They were comparing it to mechanical manipulation of symbols, which, in the 1940s, led to the invention of the programmable digital computer. It was based on abstract mathematical reasoning, and the machine itself, along with the ideas behind it, made a handful of scientists to begin thinking about creating an electronic brain.
For most people, the idea of artificial intelligence became known through movies, such as “The Terminator” and “The Matrix”. Machines that have the mind of their own, that are self-conscious and are able to think for themselves. The technology has evolved a lot since those movies were introduced to the public, and we are getting closer to the day when robots are going to be able to think and act the same as humans, or maybe even better. Maybe the most famous and the most advanced humanoid robot today is Sophia, developed by Hong Kong based company Hanson Robotics. But, in order for robots to know something, we have to find a way to teach them.
That’s where neural networks come in. Modeled after the architecture of neurons in the human brain, the “neurons” of the neural network take some input, use an algorithm to decide what to do with that input, and based on the results pass it to the next group of “neurons” until we get the final result. Those “neurons” are actually nodes and a lot of them. They are divided into layers. The input layer, multiple hidden layers, and an output layer. Every node in one layer is connected to every node in the next layer. And each connection has its value or weight. The bigger the number the bigger the effect it has on the system. Once the data gets through the entire neural network, the output layer shows the response to the data given in the begging. In order for a machine to learn something, it needs a tremendous amount of data. While learning, the machine’s output is compared to expected results. If they are the same the machine is validated, if not, mathematical equations and weights of the nodes are adjusted, and the machine is tested again.
AI is already present today. It is used in healthcare to detect diseases at early stages, for medical image processing, for precision treatments or cost reduction. It is used in the automotive industry with self-driving cars, not only driving the car without the assistance of a driver but being able to detect various hazards on the road. It is used in finance and economics, government, military, advertising, art, and video games. Making non-player characters master the nuances of the game, that take human players a long time to master, very quickly. Even though artificial intelligence is helping humans in a lot of cases, there are also a lot of concerns and newly found philosophical and ethical questions. How to deal with the intelligence that works like the one of a human, but is not? Does the machine deserve the same rights as human if it can think like one? Can a machine cause harm intentionally? And many more… Stephen Hawking, Bill Gates and Elon Musk, arguably some of the greatest minds of our time, although in support of AI, have expressed their concerns for it. They say that once the AI is fully developed, it will be able to evolve on its own at a much higher rate than humans, and in terms supersede us.
So, should we embrace AI completely, or should we keep one eye opened?
AI is definitely a powerfull tool and it could be used for improving human lives. We have experience with AI projects and we can say that AI is a powerfull tool that can be used for improving life quality. For our client, we build an image recognition tool that helps to make a list of assets using a camera and simply taking photography of asset. It was a great experience for us!