An ideal intelligent computer program can change itself to take actions that maximize its chance of success at performing a task. According to computer scientists Stuart Russell and Peter Norvig, the term ‘artificial intelligence’ is applied when a machine mimics ‘cognitive’ functions that humans associate with other human minds, such as ‘learning’ and ‘problem-solving’. Simply put, the computer ceases to be a machine that simply carries out instructions based on computer programs. It gains the ability to recognize, learn and see patterns such as a consumer’s buying behaviour as well as to analyse and solve problems such as finding the shortest way to the nearest petrol pump or restaurant on its own, with no new programs needing to be written. In essence, the computer system writes extensions to its own original programs that make it more efficient at new tasks when it processes through the data thrown at it as it is used.
Contrast this with process automation—robotic or otherwise—which takes manual tasks that do not need much learning and simply mechanizes them. This could be as simple as the scanning of invoices to be processed in an accounts payable system. All the programmer has to do is define where fields such as amount due and payment address show up on the invoice, and then program the system to look in those particular spots to find this information. This step then becomes automated and removes the need for a manual keyboard operator to input such information on to the system, thereby displacing these keyboard operators. This kind of programming is not AI.
Also read: Meet Astro, Amazon's home robot that will follow you around
Most IT services firms selfishly blur the line between process automation and AI. Despite their best efforts at obfuscation, the line is really clear. Automation simply mechanizes routine tasks. But in AI, the computer program itself learns as it goes along, creating a database of information that it then uses rules of thumb to analyse. In a vital twist that has occurred in machine learning in the past few months, these databases themselves generate additional computer programming code as they learn more, without the need for an army of computer programmers. In AI speak, this is now often referred to as ‘deep learning’.
As AI becomes more capable, it simply is no longer considered ‘intelligent’. For example, the multilingual work I used to do over two decades ago is no longer called AI since it is now routine. So will it be with many of the programs now considered to be on the cutting edge of AI.
What is not in doubt, however, is that automation—whether routine or ‘intelligent’—will lead to seismic shifts in employment, especially in India with its armies of programmers, and much like the industrial revolution in the 1800s in the West. People spinning or weaving fabrics lost their jobs after Eli Whitney invented the cotton gin in 1794, as did many a buggy-whip maker after Karl Benz invented the automobile as we know it in 1885, and several bank cashiers after the ATM was invented. History—and classical economics—has proven time and again that a revolution such as this simply changes the nature of human work in the long term (after the excruciatingly painful short-term effects of job displacement have worked themselves out). Stop suggesting computer programming as a future profession to your children, unless you’re sure they will be genius scientists.
Simon, Tom McCarthy and others founded the AI field on the claim that human intelligence ‘can be so precisely described that a machine can be made to simulate it’. This raises questions about the ethics of creating artificial beings endowed with human-like intelligence. It is not just programmers who will lose jobs to AI, but also pilots, machinists, journalists and others. Elon Musk, himself a heavy investor in AI, says, ‘With artificial intelligence, we are summoning the demon.’
This requires a quick foray into metaphysics to debunk. While tasks, whether or not they need continuous learning, can be automated, there is one thing that a soulless machine can never do, and that is to have living consciousness. Computer programs can be taught tricks that involve the application of learning, just as apes, dogs and humans can, but learning is not intelligence. Living consciousness is the key to all true cognition. Explaining ‘consciousness’ is something that every spiritual scripture has tried to address, and the world has no shortage of such scriptures. But for now, we can make do with the simple definition that says consciousness is the simple act of knowing that you are alive.
If you doubt me, then simply ask yourself who is listening to these words as you read them to yourself. Is it your learning neurons—or some other, larger field of consciousness into which words and thoughts like these come and go and are understood? Next, go ahead and ask yourself whether you will lose your job to AI—and watch your mind (your learning repository database) respond. Both the question and your answer to it, whether driven by the limbic response of fear or by the intelligent reasoning of your database, are perceived by the real living consciousness in you. And unlike learning, it is this consciousness that is the root of true cognition. Only sentient beings have it. Even apes and dogs have it, though to a lesser degree than us; a computer program automatically generating lines of code so that it can make itself more efficient does not.
Excerpted with permission from ‘Techproof Me, The Art of Mastering Ever-Changing Technology’ by A. Siddharth Pai and published by Penguin Random House India. A. Siddharth Pai is co-founder of Siana Capital and a venture capital fund manager for deep-science and deep-technology start-ups that ideally have social impact.
Also read: Tesla's robot will make physical work a choice: Elon Musk