It isn’t hard to imagine a situation where your best friend, or anyone around you for that matter, is not human. Not in the sense that your best friend is your pet or something. But that people around you, are cyborgs, or even fully intelligent robots. Seems farfetched? It shouldn’t. We’ve seen enough movies on this now, haven’t we? Transcendence talked about the The Singularity in pretty direct terms.

Thing is – reality and fiction are converging. We are moving towards a point that is referred to as The Singularity.

What does The Singularity mean?

Singularity assumes many synonyms in accordance with the context. In cosmology for example, a singularity is the centre of a black hole, where the laws of physics break down. Singularity thus refers to anything where normal rules no longer apply. In the case of computer science, or artificial intelligence, a singularity is a point when we would achieve human intelligence (or even better, super intelligence) on computers.

When I watched Terminator, the first thought that hit me was that this was fiction – pure fantasy of a creative mind. But the as days went by, I grew up to realize that science fiction would turn out to be reality in due course of time. And now, we’re closer than ever.

Related: Where is Artificial Intelligence headed?

yaabot_singularity_graph

The rate of technological innovation is not a linear graph, it is a rather an exponential curve. This phenomenon was called by futurist Ray Kurzweil as the Law of Accelerating Returns. We’ve taken some images from WaitButWhy.com (who have a great article on ASI themselves) to explain ourselves.

yaabot_singularity_graph2

It can be seen that though initially the graph takes up a linear path, there is a point where it assumes an exponential path.

How do you grade Artificial Intelligence?

Artificial intelligence can be graded as:

ANI: (Artificial Narrow Intelligence) Is the level of artificial intellgience we have achieved currently. It is also known as ‘weak AI’. These systems can perform excel only in a particular field (In a specific function). Like Google Now, or Siri, or Cortana – these ‘personal assistants’ excel is assisting you in your daily life, with mails, messages, reminders and the like.

AGI: (Artificial General Intelligence) This is the next stage wherein computers equal human intelligence; they are able to perform all human skills in thinking, reasoning and other processes of a human mind. This is also called ‘Strong AI’. Like robots, for example, like the NS-5 in Will Smith’s I, Robot.

ASI: (ArtificialSuper Intelligence) This stage is really creepy. ASI is what we’re going to call computers that turn out to be more intelligent than humans. The level of intelligence these machines will run on is beyond human comprehension. ASI will be able to solve problems considered impossible for us, within seconds.

Why aren’t we there yet?

It is extremely easy for a computer to handle complex calculations involving integrals and differentials, or atomic physics or whatever. But at the same time, it is difficult for a computer to understand and observe all the crazy phenomenon happening around it and have the ability to comprehend its surroundings like humans do.

Related: Is there an end to Moore’s law?

Bill Gates put it very nicely when he said the human brain is an extremely slow computer, running a very beautiful algorithm – the ability to question things, and possessing sentience itself. Our computers are extremely fast, but lack the algorithm. It will soon be possible to run our very beautiful algorithm on extremely fast computers made of silicon.

Before we get down to figuring out how to replicate the algorithm that gives humans intelligence – we’re going to need to speed up our computers. Because they aren’t exactly fast enough yet.

yaabot_3

Currently we need to figure out how to increase the computational power of our current systems. Hardware has to be made smaller and more efficient to process large amount of data. Once when we figure that out- achieving human percetion, emotions and understanding in a computer would be a series of developments that would be exponential resulting in singularity.

However, given the pace of technological growth – there’s a specific date futurologists have predicted for the singularity – 2035. If that is proved correct, we will be part of a very privileged generation – a generation that will witness the birth of ASI – and the world will change overnight.

Are we ready for the singularity?

But the question is, are we ready to handle the singularity when it arrives? What are the implications if we create computers that are more intelligent than us? Would they agree to work side by side for our and their development or would they treat us as inferior creatures that share this world. What if they control everything? It could take minutes, if not seconds, for ASI to figure out solutions to problems that have dodged us for decades – world hunger, poverty, cancer. Everything. 

Of course, there are other implications too. Who’s to ensure that ASI will only work to cure cancer, and not nuclear bomb the entire human race? A very stringent sent of regulations will be needed to ensure ASI does no harm. Already, many reputed individuals (like Elon Musk, Stephen Hawking) are displaying heightened concern over the potential use of AI for purposes detrimental to humanity. Robots will no longer be mechanical instruments that carry out functions in a monotonous manner. They would in course of time have the vitality and vigor of human intelligence and emotions.

Ray Kurzeil says all of us could be robots soon. Maybe it is the destiny of intelligent civilizations to progress to an electronic life-form, making biology only a phase in the life of civilizations. True singularity.

Emotions are a complex biological development that has been achieved through billions of years of evolution. Will we be able to create equivalent delicate software that handles all the nuances of human emotions and feelings? It is like a jigsaw puzzle where if a few pieces fall in place the rest would be in their respective places in no time. When a breakthrough happens, we would be able to crack the rest to achieve human or rather human intelligence on computers.   

Too much to process? A few decades down the line, we will have computers to ponder these questions themselves, further eliminating human interference on the world.