Martin Ford is the founder of a Silicon Valley software firm and the author of Rise of the Robots: Technology and the Threat of a Jobless Future. Geoff Colvin is senior editor at large at Fortune magazine and author of Humans Are Underrated: What High Achievers Know That Brilliant Machines Never Will.
Martin Ford: To understand why today’s information technology could have a much more dramatic impact on employment than anything we’ve seen before, it’s best to begin by considering the nature of work performed by most of our population. The reality is that a very large fraction of our workforce is engaged in activities that are on some level routine, repetitive and predictable. This is not to say that most people have jobs that are rote-repetitive, but rather that most workers face the same types of challenge again and again and that most of their actions and decisions can be predicted, based on what they have done in the past.
Recent advances in robotics, artificial intelligence – and, especially, machine learning – suggest that much of this predictable work will be susceptible to automation over the next decade or two. No one is arguing that all jobs will be automated or that individuals with exceptional talent or high levels of creativity will be threatened in the near future. However, for average workers engaged in more routine, predictable occupations, the impact could be quite dramatic.
While there will doubtless be many calls for improving education and retraining opportunities, it is, I think, unrealistic to expect that the bulk of our workforce can somehow be trained to take on roles that are beyond the reach of technology. While progress will certainly create new opportunities for those with the proper capability and training, it seems very unlikely that there will be enough of these new positions to absorb all the workers displaced from more predictable work, even if most workers are able to re-educate themselves successfully.
Geoff Colvin: Martin, I’m with you completely on the awesome abilities of advancing technology and I agree that most people have no conception of what tomorrow’s technology will be able to do. It indeed seems possible that, for the first time since the dawn of the Industrial Revolution, technology will eliminate jobs faster than it creates new ones.
But I’m not quite ready to bet on it. Two factors in particular demand consideration. One is that while it’s easy to identify the jobs that advancing technology will eliminate next, it has been practically impossible to imagine the new jobs that technology will create. So we overweight the reality of the disappearing jobs and underweight the reality of the new ones. Even after the worldwide web first made the internet accessible to everyone, no one predicted jobs for search engine optimisers, mobile app developers, social media managers and countless other jobs of today. We can be confident this pattern will continue. Of course, that doesn’t mean the new jobs will be numerous enough to counterbalance the disappearing ones, but it gives us reason to be humble in our predictions.
The other factor is the way technology revalues skills. The overall number of jobs could hold up even in the face of advancing technology, but many of them may be very low-paying jobs for people who haven’t improved their skills. It’s also possible – my view – that skills of deep human interaction will become far more valuable and many people will be able to prosper by bringing those skills into the evolving economy. That would be a much more hopeful future.
MF: Geoff is certainly correct that technology creates new types of jobs and also constantly changes the mix of skills required to perform those jobs. The question is whether progress will create enough new positions to absorb the victims of automation – and whether these new jobs will be accessible to people with average capability.
Despite all the hype about entirely new types of work that would have been unimaginable to previous generations, the reality is that the vast majority of our workforce is employed in traditional occupations. The most common occupations in the US are salesperson, cashier, food and beverage server, office clerk and driver. Indeed, one recent analysis found that about 90% of the US workforce is employed in occupations that existed 100 years ago. The website designers, social media marketers, mobile app developers and all the rest constitute a very small fraction of total employment and in many cases they require highly specialised, technical skills.
A great many of the traditional occupations that now employ millions of workers are going to be highly susceptible to automation over the next couple of decades. If those jobs begin to disappear and there are no new opportunities within the reach of those workers, the impact on both the fabric of society and the economy will be dramatic. Given the political challenges associated with developing a viable solution to such a disruption, I don’t think it’s too soon to begin a meaningful discussion about the risks we face and the types of policies we might employ in order to adapt our system to the new reality.
GC: To be clear, I take seriously the possibility that technology may for the first time be reducing total employment rather than increasing it and the first two chapters of my new book are all about that. And I certainly agree that it’s not too early to talk about policies for a future in which that happens. It would be a world such as we have never inhabited. The immediate challenge for individuals is making sure they’re not losers as technology transforms the economy. I think far more of them will be able to do so than even the workers themselves may believe.
Consider the idea that most of today’s jobs existed 100 years ago. Are they really the same? Some are – think of waiters, for example – but many clearly are not. For instance, the number of bank tellers in the US held steady from 1980 to 2010 despite the onslaught of automated teller machines. How could that be? MIT labour economist David Autor observed in a recent article that it happened because tellers gave up “routine cash-handling tasks” and moved to “relationship banking… forging relationships with customers and introducing them to additional bank services”. They’re still called tellers in the official statistics, but their work has been entirely transformed.
And note how. Instead of being human machines behind a counter, they’re now sitting at desks building relationships. The required skills are completely different, and they’re the skills of human interaction. For tellers and many others across the economy, these deeply human abilities are increasingly the high-value skills that will separate the economy’s winners from the losers.
MF: I think Geoff is certainly correct that individuals can and should do everything possible to adapt to the changes brought on by technology. I fear that to a significant extent, however, this may turn out to be a zero-sum game: where one person succeeds and gets the job, someone else will be out of luck. Therefore, it’s important to realise that advice directed at individuals is quite different from a discussion about what we should do as a society.
There are certainly policies that can help to some degree. We ought to ensure that effective education is accessible and affordable for those who will need to adapt to rapid change. In the US, for example, we have major problems with huge amounts of student debt, as well as for-profit schools that often prey on the most vulnerable.
Technology does indeed transform the nature of jobs, but it often does so in unpredictable ways. The conventional view has been that progress results in the automation of lower-skill jobs while creating more opportunity for those with higher education levels (economists call this “skill-biased technical change” or SBTC). Recent evidence calls this idea into question, however. Most of the jobs being created are low wage, and technology often has a “de-skilling” effect. For example, cashiers used to have to enter each price manually, doing so with speed and accuracy was a skilled undertaking. Barcode scanners have changed the picture dramatically.
I’m not sure that Geoff’s (and David Autor’s) observation that the number of bank tellers in the US held steady from 1980 to 2010 is really such a strong argument against the impact of automation. In 1980, the US population was 227 million. In 2010, it was 309 million. I suspect that those 80 million extra people likely did quite a lot of banking – and ATMs were handling it.
GC: Actually, my observation that the number of bank tellers has held more or less steady (it actually increased by about 10%) is an argument for the impact of automation.
The larger point, on which I think Martin and I would agree, is that there will be winners and losers as technology transforms the world of work. The public policy response should include helping today’s young people become winners. How? An advisory group of top British educators and CEOs last year recommended changes to secondary education in the UK. They concluded that “empathy and other interpersonal skills are as important as proficiency in English and mathematics in ensuring young people’s employment prospects”. The group urged that these skills be taught to all secondary students “but with the process of learning these starting much earlier in school life”. The competencies “should be embedded throughout the curriculum”. That is, everyone will need these skills.
Could this turn out to be a zero-sum game, as Martin fears, with a loser for each winner? I fear it also. That has never happened before economy-wide, but today we’re in uncharted territory. We shouldn’t let the uncertainties stop us from trying to help as many people as possible become more valuable as the economy evolves.
The Rise of the Robots by Martin Ford is published by Oneworld (£18.99). Click here to order a copy for £15.19. Humans Are Underrated by Geoff Colvin is published by Nicholas Brearley (18.00). Click here to order a copy for £14.40
guardian.co.uk © Guardian News and Media Limited 2010