Search This Blog

Tuesday, July 4, 2017

Of Cattle, Cognitive Systems and Creation - Part 1

Sub-title: Artificial Intelligence is a hugely powerful next step in human evolution. History teaches us how to control such power and ensure its use for human well-being.

As per popular history, somewhere between 10,000 BC and 4,000 BC, transitioning from the Neolithic Age to the Bronze Age, we humans evolved from nomadic groups of hunter-gatherers into full-fledged agrarian economies. Among other things, we had an insight that animals, which were all around us then, have strengths and skills that can be great assets to getting our own works done. We developed “technologies” to domesticate, tame, train, rear and breed cattle for various purposes. Whether it is dogs in Alaska, elephants in south India or horses in the American West, leveraging local animals to get things done became a major factor in the survival and growth of us humans. Animal-less productivity was nearly unthinkable then. Since then, we have undergone two more major evolutionary steps as a race - the industrial revolution and the IT revolution. Nevertheless, leveraging animals for human well-being (other than as food) continues to be practiced significantly.

Cattle provide an excellent perspective to think about the current Artificial Intelligence (AI) revolution. But before we go there, let us first get on the same page on what the AI revolution is. In case you were busy cave-dwelling in the past decade, let me be the one to break to you that our world is being completely reshaped by AI. Clerk-less offices, driver-less roads, teacher-less schools, soldier-less wars and doctor-less surgeries are no longer science fiction. In the coming decade, intelligent machines and robots will save millions of human and animal lives, but they will also steal millions of human jobs. Companies offering AI-based products and services will earn trillions of dollars. Cities, institutions and homes will look completely different in our life time, thanks to AI.   

AI is a technology area that aspires to build computing systems that exhibit what humans can accept as “intelligence”. The dream of creating intelligent programs and systems has been active for several decades now. Many connected technology areas like Big Data, Analytics, Robotics, Machine Learning and Cognitive Computing assist in realizing the AI Revolution. It is known that thousands of mechanical, electronic, chemical systems outperform manual effort, in many areas. When the ingredient of intelligence gets added to these systems through AI, what we get (what we are getting) are truly next-generation machines that do complex activities far more quickly, cheaply, accurately, reliably and robustly than human beings. A sub-area of AI, Machine Learning, enables these machines to learn from their mistakes and from human feedback, to improve over time at what they do, just like humans. In short, AI makes it possible to create perfect, intelligent workers in many areas.

Naturally, this is highly scary. At once, our very survival seems to be threatened and mind gets clouded by many questions - What will happen to the human edifices and societies that we so carefully constructed for centuries, when AI takes over? Will humans have anything useful left to do? Or will all work be effortlessly done by AI-enabled systems, forcing us to become couch potatoes and eventually go extinct? Once machines become intelligent, how can we be sure that we can continue to control them to keep serving our interests and mandates? What if they revolt? Is it possible for quirky, emotional humans to stand their ground and fight against perfect, ruthless machines? Can intelligent machines build consensus among themselves, to become perfectly coordinated nightmares for humans? Can the intelligence of machines grow so much that humans become vulnerable disposables? There are many such questions, and they must be asked and debated. These questions are valid enough that some of the brightest minds of our times, like Stephen Hawking and Elon Musk, have alerted humanity on the not-so-distant future dangers of AI systems.

However, I am a die-hard optimist. I choose to hold a diametrically different view than the above venerable minds. My 27 years in Artificial Intelligence and Machine Learning have convinced me that AI systems are nothing but humanity’s new cattle. Like our Neolithic brethren woke up to the possibility of leveraging the animal kingdom around them, we are waking up to leveraging a new-found kingdom of intelligent machines. This is a kingdom we will breed. We can, and we will, learn to leverage this kingdom in the right way, to advance human well-being, to thrive and prosper as human race, and to strengthen the human spirit. In the initial stages, there will be challenges and casualties, just like there are in initial horse-raring and elephant-training times. The more powerful the beast is, the more potentially useful it is, but the harder it is to tame, train and use. But I believe there are solid, irrefutable reasons to prove that we humans can tame the AI beasts.

Through this article series, I hope to convince you that we should not shy away from pursuing or building AI-based machines; on the contrary we should build them with a sense of passion, urgency and discretion. In parallel, we must put in place specific mechanisms and measures, to help create a future where the supremacy of humans over intelligent machines can be assured.

To be continued…

[This is a multi-part article series, which contains my personal views and ramblings about natural and artificial intelligence. Thank you for reading. Please watch this blog for updates.]