Artificial Intelligence : Should we be concerned about it!!

Intelligence is the biggest contributing factor in human civilisation. Intelligence is what gave rise to technology which facilitates the creation of various products to automate and simplify human life. Artificial intelligence is the amplified form of human intelligence intended towards flourishing civilization like never before. Artificial Intelligence though often portrayed in science fictions having human-like characteristics, it can actually encompass everything starting with Google’s search algorithm to automated weapons, from SIRI to IBM’s Watson, from facial recognition to self-driving cars and so on.

Artificial intelligence as of now is being considered as narrow or weak AI as, it is designed to perform some small specified tasks for humans like solving equations, responding to voice commands etc. However, in the long run, researchers are on to creating General AI or strong AI that would be able to outperform human beings in almost all cognitive tasks.

Why safety is a concern with AI?

The concern with AI is when in the long run, the quest for designing a Strong AI will succeed and AI will become better and smarter than humans (if, in case), such a system might be able to set goals of it’s own, denying human command and performing tasks on it’s own leaving human intelligence far far behind. Some argue that AI might be able to invent new technologies with super intelligence which may help humankind in eradicating diseases, poverty, war and deal efficiently with natural hazards so, creating a strong AI would be the biggest invention in the human history. Alternately some other researchers are of the view that it might also be the last invention in human history unless, we are able to align our goals with that of the AI before it becomes super intelligent and starts setting its own goals.

There are many holding the view that achieving a super intelligence will never be possible and, there are also others who strongly believe that a strong AI is pretty much an attainable goal. Also, there are others who believe in both the possibilities and also are concerned about the potential harm (intentional/unintentional) from the strong AI and prepare to deal with the same.

Is AI really that dangerous!!

Most of the researchers agree to the view that a super intelligent strong AI is quite unlikely to have human emotions and it’s not actually practical to expect AI to feel and understand human emotions. However, AI not understanding human emotions can be harmful alternately, AI able to understand (if) human emotions are a much bigger risk. Either way it can be programmed to perform something devastating. Take automated weapons with AI that are programmed to kill can be really dangerous in the hands of a wrong person and could lead to mass destruction. Moreover, when AI is capable of developing goals of it’s own and start refusing commands, it can act on it’s own and it will be impossible for the researchers to stop it.

This may sound like a science fiction but, not to those who have heard about the recent news on Facebook shutting down one of it’s AI systems as things got out of hand when, the chat bots developed a language of their own without human interference. Yes, according to reports of Tech Times, the Social Media giant had to shut down the AI system they were researching on when they found out that the AI bots have developed a completely new language on their own. According to the report the AI agents were initially using English to communicate within themselves but, suddenly they stopped using English and created a language of it’s own that, only the AI itself was able to understand and thereby defying the codes provided. Finally, the researchers had to shut it down in order to make it respond to the codes.

In June, while the researchers at Facebook AI Research Lab were busy working on the improvement of chatbots, they found the dialogue agents started creating a completely new language of their own. Soon after that the bots started deviating from the scripted norm and began communicating in their own language which they themselves have created without human input.

These bots were also found be incredibly crafty negotiators. After learning to negotiate, these bots attempted to improve the outcome of the negotiation using advanced strategies and machine learning. It may sound incredible but the bots actually became quite skilled at it over the time and they even began pretending to have interest in a particular item such that they can sacrifice the same at a later stage of the negotiation as faux compromise.

This may be alarming although it appears as a great leap for AI. this could be a threat to the human kind who are limited by slow biological evolution.

Concluding Note..

The FB incident is evident that how devastating AI can turn into. Many big names in Science and Technology (like Stephen Hawking, Elon Musk, Steve Wozniak and many others) and researchers have already expressed their concerns on the possible risks posed by AI. But, why this matter suddenly attracted so much attention? Why everyone is talking about it!!

This may be because, the possibility of designing a strong AI seemed to be a distant dream more, centuries or more away. However, with recent breakthroughs, getting this dream come true in near future seems quite possible, like some decades or so away. And, as AI possesses the potential to become more intelligent than any human, concerns related to safety with AI has risen.