Girish Joshi

The AI Cliff-Hanger

5 min read

Artificial Intelligence has defeated us in checkers, chess, scrabble, backgammon, and Othello. With every passing day, they are getting good at recognizing faces, understanding natural language, performing complex motions, recommending things that make us happy, and detecting spams and frauds. If you follow tech then you’d know that now they can even make phone calls. Our phones know about our patterns (thus predicting our behaviour) more than our friends. Almost every online service employees AI to make our experience better. And soon enough AI systems will be driving our cars. Meanwhile military is using AI to guide missiles better, to make drones more smart and accurate, to make robots that can survey remote areas and destroy targets. Every industry on the planet is exploring avenues for AI for there is no survival without it. Even the dinosaurs, The Oil & Gas Industry. And yet we have only started out with Artificial Narrow Intelligence, the best (and scary) part is yet to come. But before we dive deep, let’s brush on some basics.

What is a Computer?

A fancy calculator. All it knows is how to perform arithmetic and logical operations. And it’s very good at what it does and defeats humans 101%. Ask it what is 1795 x 89453 and it will blurt out the answer before you can even blink your eye. Ask it to differentiate between apple and orange or write a novel and it will stare at you in the abyss for eternity. Turns out a computer is extremely good at things that require humans to think like calculus, solving a differential equation, or finance and extremely bad at things that humans do effortlessly like perception, motion, or speech.

What is Artificial Intelligence?

AI is making computers do all the things that humans can do and not just math. It’s our attempt to simulate the human brain. A computer with human-like capabilities will not only be efficient and fast but it will also defeat humans at 101%. Apparently, there is a funny paradox with AI; as soon as it works, nobody calls it AI anymore. So although AI is here already, it still seems like a thing of future.

Then What is Machine Learning, Data Science, Deep Learning?

They all are the subsets of the superset AI. But we’ll not talk about that today. There is enough literature on the internet to kickstart your adrenaline, go explore once we’re done.

Now let’s begin.

All the AI examples I mentioned at the start have one thing in common, they all are good at one specific task and that’s it. Humans aren’t like that. The best chess player can drive a car equally well and can also sing for recreation. The AI systems that are good at only one (or few) specific tasks belongs to class – 1 of AI. This is Artificial Narrow Intelligence or ANI. This is the only class we have mastered until now, or we chose to think that way.

Now for a moment try to appreciate how incredibly powerful and efficient our brains are. Have you ever marvelled at seeing a baby learn the first word or take the first step? Wonder a computer program doing that, learning on its own, and not just one thing, but everything, and learning it faster and better than us. But right now only the worlds fastest supercomputer matches our brain in calculations-per-second, and we know for fact how humongous supercomputers are and how much energy they consume. And no one really knows how to write such a program.

The class – 2 of AI, one we still are figuring out how to create is called, Artificial General Intelligence, or AGI. AGI will be as good as a human, it would be able to plan ahead, strategize, self-learn, and do all the things we humans can do with intelligence but only better and without fatigue. Now from here on it may start to sound like another sci-fi movie plot but most of the experts believe that we’ll reach AGI by the end of this century, and that’s the most pessimistic estimate I can give you. Optimistically AGI would be reached before 2050.

If AGI is reached, the life as we know it will forever be changed. We’ll have a digital species that is as smart as us, that is never tired, that is conscious all the time, that learns faster than us. And once AGI is reached, it will be only a span of time in which class-3 of AI will be reached. Artificial Super Intelligence, or ASI. ASI will be way intelligent than humans. To bring things in perspective, what ants are to humans that humans will be to ASI.

Imagine if we mere mortal humans could invent Wi-Fi, create the Internet, and could cure Cancer. What would an immortal machine with human-like capabilities do? The solution to world hunger, energy crisis, climate change, or any problem would be with us. Solution to ageing would be there, immortality can be reached. Obviously for that to happen, nanotechnology has to keep pace with advancements too.

But that’s the bright side. On the contrary, all the evil that man has done will also be amplified. The notion of good or evil is humanly devised. Applying it to non-animate things is called anthropomorphizing. ASI’s are not supposed to understand it. Morals and ethics will mean nothing to them. Hence an ASI would do anything to accomplish its goals, even if it means killing all humans and colonizing the entire universe. Elon Musk calls it “summoning the demon”? Sure his concerns are valid. Creating something smarter, (way smarter) then ourselves can be a Darwinian error. The excitement right now is same as when single-celled organisms were turning into the multi-celled organism. We are amoebas and we can’t figure out what we’re creating. No wonder why some call it our last invention.

Is this a path to immortality or extinction? Only the time will tell. Now is the best time to be alive.

Read about The AI Revolution at Wait But Why to have your mind blown.



Leave a Comment

Your email address will not be published. Required fields are marked *