Employment in the time of AI

Artificial Intelligence (#AI) has stormed into public consciousness. There is euphoria about what it can do for everybody, mixed with fear about what it may do to individuals. Young people have more cause for concern. How will #employment change? Will #jobs eliminated by AI be replaced with new jobs? Who will be rewarded? Who will be discarded?

While the future in its entirety is unknown, some things can be predicted using data. I offer below some numbers and implications.

#India has an unmatched number of young people. No other country is close. Below age 25, the population is about 25 million per year. The number of people between ages exactly sixteen and seventeen is about 25 million. Between exactly seventeen and eighteen? Another 25 million. And so on. For context, all of Australia has about that many people. We have, roughly speaking, an Australia of fourteen-year-olds, another Australia of fifteen-year-olds, and so on. It has begun to decrease slightly among the very youngest children, but not enough to change the main picture.

Now consider #layoffs, which are a common part of modern employment. 45 is a bad age to be laid off. If one is laid off at age 25, recovery is easy. At 35, one might compromise slightly on #salary or location, but many years of career-building remain. Even if one is forced to move a little lower, there is time to accept it and, with luck, find contentment. From the other end of the career span, being laid off at 65 is merely #retirement. Being laid off at 55 may involve taking a new job in a different domain at even half the pay, but the damage is limited because one may already own a home, one’s children may have grown up, and one may have accumulated savings. Age 45 is in the uncomfortable middle. That is when many professionals have small children, unpaid home #loans, #spending habits that are harder to change, and the first signs of physical #aging.

Next, consider people presently in #college. A typical age is 20. As noted above, we have 25 million 20-year-olds, although most of them are not enrolled in good colleges. Behind them we have about 500 million who are younger.

I am 56 years old now. Over the next 25 years, many of my cohort will die. But not those who are presently of age 20. After 25 years, in 2048, most of them will remain alive. Behind them there will be work-seeking people aged between 20 and 45, packed in at about 25 million per year. That is 25 times 25, or 625 million. The 45 years olds of 2048 may suffer layoffs if even a small percentage of the 625 million can claim the same #skills, and do the same work for less #pay.

Admittedly, I was 45 myself once. My cohort also had younger people behind it. But we did not face such brutal consequences. Why? Because we grew up in a less equal time. Children born into #poverty had such a #handicap that better-placed ones like me faced lower #competition. The misfortune of many was the good fortune of some, myself included.

But the 625 million people I describe above will be different. They will have grown up in a world where almost everybody had a mobile #phone, #internet access was cheap, and the #knowledge of the world was free. Those who wanted to learn, and were prepared to work hard, did not need admission to a good #university. If they wanted to #learn, they could.

Skeptics may say that, internet or not, most people never learn much. True, but not all 625 million need to learn much. If even one single million among them can learn useful things, and can prove they have learnt them, then the Indian job market will feel it.

To understand, look at learning over time. In old times, the student had to travel far and spend time with the original #teacher. The printing press democratized matters. Students took mass-produced #books to their homes, and the #lecture hall was born. The lecturer hit the high points of the book, offered philosophical detail or clearer interpretation, and helped hundreds of students at once. Then came #video lectures on the internet, and the lecture hall itself went with students to their homes.

Was the university made redundant, or reduced merely to the evaluation of answer scripts and the granting of #degrees? Perhaps not. Online lectures are #suboptimal. Live lectures are responsive because good lecturers read faces and make small adjustments, while internet lectures are static. Moreover, the shared experience of sitting in a #classroom with friends has value for learning and beyond.

But we are not talking about all. We are talking about just a few. Maybe just one million out of 625 million. A million which did not have access to the lecture hall to begin with, but had high motivation and drive. For that million, internet lectures could make a big difference.

And finally, come to AI. If AI can teach in a responsive way, then internet lectures may lose their static feel. If AI can also evaluate candidates in an evolving and responsive way, then the university degree might lose its permanent stature, too. After all, a reliable recent #certification may be more valued in a 35-year-old prospective #employee than their degree from 13 years ago. With AI, both teaching and evaluation could become simultaneously cheap and #scalable.

Which makes one ask, why merely one million out of 625 million? Why not one percent, or 6.25 million? Who says that AI cannot properly train, evaluate, and certify one percent of the population? And if it does, those paradigm-changing 6.25 million workers, rising in a hungry tide that laps at the feet of the 45-year-olds of 2048, may cause layoffs in big numbers indeed.

I tell students, do not worry about work that AI will take away from you. Worry about the work it will take away from your teachers! If AI teaches certain skills to millions, how much pay will those same skills fetch you? Given a choice, do not choose a higher salary today. Choose, instead, work that AI might find difficult to teach tomorrow. And while I cannot say definitively what that “safe” work will be, at least you have time to think and plan. An imperfect plan may be better than no plan at all.

To those who want greater #impact, social change, wider #opportunity for all, as well as a #business theme for the next 25 years, I suggest, use AI to teach India how to do useful things. These useful things are not mysterious. They are the same things all good colleges teach. #Logic, #mathematics, #science, #technology, #computing, #accounting, #psychology, #design, #speaking, #writing, and so on. Use AI to take good teaching and evaluation to the people. There, even as you contribute to greater social equality, you will find 625 million opportunities.