My son is at that age when he needs to decide what he is going to do next. For the past fourteen years he has ratcheted through the school system without having to make a big decision. Now he has to decide whether to go to university and what to study, with an eye on where that will lead.
The decision comes against a backdrop of profound change for ultimate employment prospects. He will be thrust into the workplace just as a new industrial revolution is taking shape, built around artificial intelligence. It is not a stretch to assume that machine learning and robotics will change both the nature and the relevance of many lines of work that a graduate of old may have aspired to.
Many estimates have been put out there for the amount of job destruction that new technologies are likely to wreak. The most extreme estimate, proposed by Frey and Osborne, suggests that 50% of US employment is at high risk of automation. Other estimates, which take into account economic viability as well as technical viability, suggest lower figures. A recent comprehensive study by the OECD puts the proportion of jobs at risk from automation across 32 countries at 14%. This seems comparable to the experience in prior technological revolutions.
There are two differences though between this unfolding revolution and those of the past. One is that, like everything else about it, it is happening more quickly. The second is that whereas previous waves have targeted manual jobs, this one is aimed at jobs with a cognitive element. Artificial intelligence is increasingly able to replicate hitherto uniquely human skills like intuition; it also has scope to outperform humans along dimensions of consistency, connectivity and updatability.
This second point is critical for a boy standing on the edge of adulthood because everything we have prepared him for reflects a continuation of the historic trend. Fighting the last battle, we have armed his generation with an education geared towards cognitive work. Indeed, in the past twenty years we have doubled down. When I went to university only 19% of people my age in the UK did; by 2016 the figure was 46%.
The Frey and Osborne and OECD studies analyse the kinds of jobs that are at risk. None are immune. Even the professions, long vaunted as a hedge against both economic cyclicality and structural shifts, are exposed. Research by Richard Susskind suggests that traditional lawyers will in large part be ‘replaced by advanced systems, or by less costly workers supported by technology or standard processes, or by people armed with online self-help tools’.
The hollowing out of industries even in the so-called knowledge parts of the economy is something with which I am familiar. I started work in the mid 1990’s as an equity research analyst for an investment bank. At it peak at the turn of the millennium there were many thousands of people doing what I did. Yet over the following years the number shrank. One measure, of capital consumed by the industry globally, has fallen 50% from its peak. Looking at LinkedIn profiles of former colleagues reveals that less than a quarter of the people I worked alongside 10–15 years ago are still active in the industry. In my case it wasn’t technology that sparked the disruption — it was economics and regulation. Although since I left it is fair to say that technology has disrupted many of the low-level tasks I performed. I was lucky. My career peak coincided with the industry peak and I was able to switch into an alternative role that was less subject to the same level of structural decline.
One of the similarities with previous technological revolutions is that new jobs will likely be created. Since the beginning of the industrial revolution, for every job lost to a machine at least one new job was created. When I was graduating from university, there was no market for social media managers, app developers or cloud architects, yet they emerged to fill the gaps left behind by FX traders for example.
The problem for a graduate today is that as we stand on the cusp of this new revolution, they don’t know where those opportunities may come from. And as in my generation but unlike in prior generations, skills are less easily transferable as workers specialise, which makes it harder to switch paths once the journey has begun. Indeed LinkedIn confirms that none of my former colleagues in equity research ever retrained as social media managers.
So where does all this leave my son?
There are broadly three responses. The first is: learn to control the machines. This is the ‘plastics’ cry of our age — learn to write code. Actually, the response is broader than that since there are many other inputs that are required to control the machines. One is the ethical overlay. It is no coincidence that Google searches of ‘the trolley problem’ began to increase in 2014 as the outlook for autonomous driving became more rosy. Perhaps the ancient job of a philosopher will make a comeback.
The second response is to do what the machines can’t. As some commentators have noted, machines first substituted our hands, now they are substituting our brains, but they are yet to substitute our hearts. One of the things they can’t do is empathise. It is notable that the US Bureau of Labour Statistics projects that a third of all new jobs over the next ten years will be in the healthcare and social assistance sectors.
The third response is to go back to artisanal ways of old. In consumer markets there is a perennial cycle whereby products start off exclusive and expensive, then become commoditised and cheap, finally as people become less price sensitive, artisanal versions appear. Think of craft beer. As technologies speed up that cycle and at the same time increase the reach of the artisanal versions, there is scope for mass creativity to be unleashed by a new artisanal class of workers.
No doubt, like all teens, my son will carve his own path. What is certain is that, unless he chooses to become a philosopher, a nurse or a brewer, it is likely to be a lot bumpier than career paths of old.
Originally published at https://www.linkedin.com.