In September 2013, two researchers from Oxford, Carl Frey and Michael Osborne, published a paper called, “The Future of Employment: How Susceptible are Jobs to Computerisation?1

In the paper, the researchers found that among the top 10 jobs most at risk of being replaced by automation include Mathematical Technicians, Tax Preparers, and Insurance Underwriters2. These are not blue collar jobs that are being replaced; in the past, where machines tended to replace human brawn, they may now replace the human brain.

This is especially true in a global environment that is spending more and more money on research in Artificial Intelligence (“AI”). In China, AI has become a strategic initiative of the government, with its desire to establish China as the world leader in AI by 2030, ahead of the United States.

Funding for AI reflects this shift; in 2017, 48% of the total equity funding of AI start-ups came from China, compared to 38% from the United States. This is a substantive increase from the 11.3% of global funding that China made in 2016.

The nature of AI technology today is to create machines that learn from themselves. The development of AI over the past few decades has been along two approaches. To illustrate these two approaches, let us suppose we wanted to identify a picture of a cat.  The first approach would be for programmers to code an algorithm that attempts to describe a cat from a top-down approach (four legs, whiskers, pointy ears, and so on). The second approach would be to feed as many images of cats as possible into a machine and have the machine learn for itself what a cat looks like based on all that data.

While data limitations in prior decades curtailed the prominence of the second approach, in recent years, given the unprecedented expansion of big data, the second approach has taken precedence. And so, we now have machines which are designed to learn on their own and, in many ways, learn in a ‘black box’ manner which is near impossible for humans to reverse engineer.

This approach has made AI particularly strong in narrow fields – the AI of today is designed to answer very specific questions such as, “What are the consumer patterns in telecommunications?” or, “How do I try to generate alpha from my investments returns?” or, “How do I optimally match people who want rides with people who offer rides?”

And so, if AI will continue to be a rapidly evolving technology globally, and if AI is becoming ever more efficient and effective in solving the specific questions for which it is designed by learning from itself, where do humans fit in?

Perhaps a cue could be from the Oxford researchers. In that same paper, they find that the jobs that are least likely to be replaced by automation include mental health workers, social workers, doctors, psychologists and teachers. This is instructive. The common denominator between these jobs is a strong relationship with people or rather, jobs that require a deep sense of humanity.

As such, any education policy may do well in answering the following question – “If AI can basically take over a whole bunch of tasks – even tasks that are ‘high-skilled’ – what advantage do humans have over AI?” It, therefore, stands to reason that what an education policy in an age of AI really comes down to is the deep focus on the greatest strength and advantage of humans over AI, which is the ability to be human. This means education that builds empathy, kindness, understanding and acceptance of others, however different they may be to us.

If we follow this logic, and we accept the fact that machines now learn from themselves, the idea that we should all learn to be master coders is very flawed. The machines will be able to write code and algorithms in a more effective way than humans. Some may argue, “Okay, that’s all well and good, but we still need to understand how the machine is doing, what it’s doing.” But, as an MIT Technology Review article pointed out 3, “No one really knows how the most advanced algorithms do what they do.” So, if we cannot fully understand what’s in the ‘black box’, why do we all need to try doing so as opposed to letting experts do that? 

To be clear, I’m not saying, “Please don’t learn how to code.” Do it if you feel it could be interesting for you or if you are passionate about it. We will need coders. But the idea that coding is going to be the ‘learn-this-at-all-costs’ skill that we all must have is, in my view, hugely overrated.

Rather, where we should aspire towards is to become more human. This means behaving in a more co-operative, altruistic and pro-social manner which, truth be told, has been the foundation of our species’ cultural evolution and, indeed, evolutionary biology.

Where do public schools fit in all of this? The nature of our economic system means that, without active redistribution, the rich get richer, and the poor stay poor in nearly all aspects. This is true even of education. If private schools provide higher quality education than public schools, and if private schools are prohibitively expensive to the bulk of the population, then only the children of high-income parents get the opportunity to attend private schools. This then sets them even further ahead on the lottery of birth which they then pass on to their children and so on.

But here’s the danger with that – the diversity of backgrounds and walks of life that is so key to our ability to connect and be human with one another may be missing in such schools. Someone who attended such a school once told me, “I spent many years in a public school before transferring to a private school because we moved. The kids from the private school are basically the same people.”

In other words, they may be diverse in nationality or hometowns or ethnicity, but they all basically came from similar walks of life. To be clear, it is unfair to generalize this to all private schools, but given the barrier to entry (read: Money) of private schools, what that person told me is not entirely a stretch.

And this is where public schools are so important. With some active interventions here and there to ensure a diversely-represented school – particularly in terms of household incomes – public schools remain the best way for children to learn how to interact and communicate with people from all walks of life and to learn that, ultimately, we are more the same than we are different.

I personally believe that the most important thing I learned at school was actually building social relationships, less so things like History, Additional Mathematics and so on, all of which, to an extent, I could have read on my own. But I couldn’t have known what it was like to make friends, fight with others, fall in love, have my heart broken, and so on, if I was just on my own.

In the same way, we need other humans – especially humans that are very different from us – to be more human. And it is in that way that we can try to achieve a more cohesive future for ourselves in a world of AI. We need an education that focuses on fulfilling our potential to be human, and we need that education to be centred on public schools where money, inherited over time, is no barrier to entry.

 

Nick Khaw is an economist at Khazanah Nasional. All views stated here are his own

References:

1. Frey, C. and Osborne, M. (2013). “The Future of Employment: How Susceptible are Jobs to Computerisation?” Oxford Martin School Working Paper no. 7.

2. This paper is not without its flaws. This essay is not the right avenue to critique the paper in detail, but it is worth acknowledging that computerization affects ‘tasks’ more so than ‘jobs’. A ‘job’ is a collection of several ‘tasks’, some of which are more at risk to automation than others.

3. https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/