Artificial Intelligence (AI) is being considered as the next technology evolution, which would disrupt a large chunk of jobs performed by human beings. Learning machines, which are being designed, are likely to learn faster than human beings. Consequently, they could even turn out to be our foes, or pose as survival threats to human beings. Not surprising though, a host of important personalities, such as Bill Gates of Microsoft, Elon Musk of Tesla, and Stephen Hawking, the famous physicist of our times, have warned us about the serious consequences that AI would cause to the human race.
The last few years have witnessed lot of work, and implementation in the automation space, particularly in Artificial Intelligence. Without appropriate regulations in place, there could be valid reasons for such fears to be expressed. Think of a scenario where AI is in wrong hands, and is being used with evil intent!!
The stage of AI research today is far below the level that could cause robots to dominate over human beings. Essentially, AI is based on code, or algorithm written by developers. Therefore, the focus of AI being used is limited and focused to specific jobs in hand. This means, the robots with AI are still stupid. They are totally controlled by scientists who design their intelligence. Usage of Artificial Intelligence today is limited to specific activities, which are repetitive in nature. Robots that are designed based on AI focus on a narrow scope of work- they possess Artificial Narrow Intelligence (ANI). Human intelligence, on the contrary, is manifested in multiple tasks, and encompasses a wide assortment of activities that cannot be easily replicable.
Examples of AI could be a driverless car, or IBM Watson’s Jeopardy supercomputer, which could potentially beat the world chess champion, or self-learning bots. How do you think that these objects work? Simple…. They rely on a large set of data – Big Data, a huge number of “ifs and then” programming, i.e. a large number of algorithms designed by human beings, NLP (Natural Language Processing, for their language skills, if any), etc.. If there is a defect, or bug in programming, the computer has to obey the master and, cannot automatically rectify itself. Therefore, the AI systems are not creative enough to build upon themselves. They still depend on their masters who design, and completely control their operations.
Like any other machine, an AI based machine has its own limitations, and could be dangerous if not controlled. Take for instance the recent “driverless car” from Tesla which met with an accident, killing the onboard human driver. Unless the algorithms are properly written in the AI processors, such costly mistakes are bound to happen.
Human beings are blessed with “common sense” that helps in taking decisions easily, whereas these fifth generation machines lack this human attribute. They are essentially “dumb” outside whatever they have “learnt”. They still have to be monitored, mentored, and controlled by human beings. Hence, fears related to apocalypse are not well founded.
As more automation is built in, and around our systems including those with AI, existing jobs would get replaced, or in some cases, updated to extend the AI based tools. Therefore, introduction of Artificial Intelligence based robots also could replace, or extend some of these jobs. Nevertheless, because of “common sense”, human beings would be better placed than robots to evolve, and take up new roles, which would add value to our society. Time and again, evolution of various technologies have taught us this lesson, and AI would be no different.
