Skip to main content

What will the digital society do to jobs? Will it usher in an age of unemployment or an age of leisure? Will we be served by robots or displaced by them? Will there be greater income inequality or less? This week: some comments on that theme, with a closing (counter-intuitive) conclusion.


Five scenarios are put forward where AI, robotics and employment are concerned.

Where are we heading? - the extremes

At the extremes, two end-points with the same basic scenario: AI and robots will replace many/most jobs, including middle-level desk-bound jobs which were not threatened by previous waves of automation. They will do this because machines can do many things that people do today much more efficiently, productively and cost-effectively.

  • Some think this means we’re heading for disaster – unemployment, underemployment and loss of income for the majority in a world that is increasingly unequal.

  • Some think it means we’re heading for nirvana. I listened recently to one of development’s most influential “thought leaders” describe the loss of jobs that’s likely here as “great”: a chance to free humanity from drudgery (with, maybe, a little risk of greater inequality).

And in between?

Three views lie in between. It’s (probably) going to be OK, these views suggest:

  • because previous waves of automation – like the Industrial Revolution – have led to greater prosperity and welfare and new types of jobs

  • because we can shift our human roles to things that we do better than machines – ‘care-giving’ and ‘creativity’ are often cited

  • because we can retrain ourselves with skills we’ll need in future – which mostly means, when governments eye policy initiatives, digital skills.

How should we think about this?

As usual, much discussion here is binary. But this debate, like many others that concern this blog, is too important to be left as a contest between optimistic technophiles and pessimistic technophobes. I’ll suggest three basic points to bear in mind when it’s discussed, comment briefly on the three views ‘in between’ above, and suggest ways in which governments and others should reconsider their approach.

  1. Jobs will be displaced

This basic fact is clear. Both extreme positions are right here: AI and robots will replace many/most jobs, including middle-level desk-bound jobs which were not threatened by previous waves of automation. They will do this because machines can do many things that people do today much more efficiently, productively and cost-effectively. The questions are to do with how fast this happens; how much it varies; and what will be its impact.

Variation between countries matters. Jobs will be displaced when it is cost-effective to substitute machines for humans, capital for labour. That will obviously happen sooner where labour’s more expensive, so it will happen more quickly in the global North than in the global South. With what effects?

  1. There’s risk of greater inequality

This risk is real and serious. ICTs tend to advantage those (countries, people) with the resources – money, skills, connections – to make effective use of them more than they advantage others, and this is likely to increase inequality.

The risk that’s most widely foreseen here is growing disparity between relatively few people at the top of earnings pyramids and the majority below, with economic and political power becoming more concentrated amongst higher earners.

It’s not a new thing, obviously – many past societies have been unequal just this way – nor is the fear of it arising from technology new either – see H.G. Wells’ novel The Time Machine (1895) for an early exploration. But it implies the need to put the risk of growing inequality at the heart of policy approaches to innovation and employment.

  1. People value work and the quality of work

The development “thought leader” that I mentioned earlier valued intellectual work much more than manual work. Everyone, he suggested, would feel better off if they did the kind of work that he did (teaching, writing) rather than (say) working down the mine. Which, perhaps, betrayed his background. For most people, “intellectual” work’s not superior to manual work, just different. Which, to my mind, is how it should be seen.

A spot of personal. The part of Britain I grew up in lost its three main employers – coalmining, shipyards, steelmaking – in the 1980s. They haven’t been replaced by "intellectual" work, but by light manufacturing, call centres and job insecurity. The only large-scale industry that’s entered local labour markets is car manufacturing, which is highly automated. Clerical work for government departments – the other relatively large employer – faces the prospect of automation now. Gig jobs and zero-hours contracts are common.

Most people don’t have comfortable lives, even in developed countries. They value work because it gives them income and strengthens their sense of worth. Job security and employment rights (such as pensions and sick leave) are highly prized. Insecurity and unemployment are feared by most people because they jeopardise so many aspects of their lives, not least security. Policymakers today have little personal experience of life on margins; they should listen more to those that do.

So what?

We really don’t know what will happen with automation and employment, though I’d say it’s always foolish to err on the side of optimism. We surely need more research on what is happening today, and more thought about how to mitigate what might happen tomorrow. Let me turn to the three viewpoints I described above as ‘in between’, and see what they imply.

The Industrial Revolution turned out well, and so will this one”

First, that notion that the Industrial Revolution turned out well, including for employment, so we needn’t worry. As I wrote last week, we should avoid a ‘Whig interpretation of technology’, the idea that innovation’s always progress.

The first, second, third industrial revolutions (steam, electricity and mass production, communications and computing), like the agricultural revolution long ago, did, it's true, create new types of work to support a growing population (and enabled greater prosperity and welfare in the longer term). But that wasn’t necessarily the case short-term, nor does it guarantee the future

In the short/medium term, for example, the first Industrial Revolution seems to have led to wage stagnation and poorer living conditions. (It’s also left us with some long-term problems – pollution, climate change, resource depletion.) That short term matters in terms of people’s lives and of political stability. Ignoring its potential downsides, on the assumption that things will get better later, isn't wise.

What’s happening today is very different, too, to what happened with the steam engine. If that precedent tells us anything, it’s that it is important to pay close attention to the impacts of technology and address not just employment impacts but the wider socio-economic changes that surround them (urbanisation/de-urbanisation, levels of income and equality, impacts on demography and the environment, physical and mental health). Policy should mean more than just waiting to find out what happens if we don't do anything.

The future’s bright; we’ll all become creatives"

What of the notion that human labour will shift from manual and repetitive tasks machines do well to ones they’re not so good at – the service sector generally, and particularly tasks that require empathy, such as care work, or “creativity” and leisure.

I’m not convinced by this. Four things to bear in mind:

  • Such jobs don’t suit everyone. Care work requires particular skills, just as much as other kinds of work. You can’t shoehorn people into looking after others anymore than you can shoehorn everyone into carpentry or lecturing.

  • They’re not as well paid as the manual and administrative work that they’d replace. Care work’s notoriously underpaid across the developing world. Restaurants and other leisure services typically pay minimum wage. Few “creatives” make much money: the occasional superstar’s atop a pyramid of people scraping by.

  • Many of the new jobs emerging now in the service sector are badly paid and insecure compared with those they are replacing. Workers in the ‘gig economy’ lack employment status and employment rights. Advocates say this gives them flexibility, which may suit some (especially the young and unattached), but it offers no security, which certainly does not suit others.  The experience of work needs to be understood over a lifetime, not just in the one-off here-and-now.

  • Supply and demand apply here too (of course). The more people look for jobs in these sectors, the more incomes will fall.

All of this requires rethinking in employment policy – recognising that the nature of jobs is changing but also addressing the ways this impacts on employment value, incomes and security.

What we must do is ensure that everyone is digitally skilled”

Many governments now recognise the need for employment policies that address the needs of an increasingly digital society. Most digital employment strategies are concerned with increasing the digital skills base – ensuring schools teach them and retraining older workers in the skills they’ll need to secure jobs in a market that demands those skills. This is entirely right and entirely appropriate, but there are two problems, on which I’ll end.

First, the digital skills that are required change very rapidly. Ten years ago, I was told by African employers that the skills of new computer science graduates were so out of date that they required retraining when they started work. Many governments today encourage schools to teach their children coding. A digital skill indeed, but how much coding will be done by humans, rather than machines, when today’s ten-year olds enter the job market? It’s going to be soft skills rather than specific skills that matter.

Second, and this is my key point, governments should think again about the implications of that notion that future jobs for humans will come in service areas like care and the creative sector. If that’s so, a jobs strategy for the digital age should not be focusing so much on digital skills, but on the non-digital skills that are required in those sectors – the things machines can’t do and humans will continue to. I know of no digital jobs strategy that is doing this right now.

Image credit: CPG_DELIVEROO_THEBAR by TaylorHerring on Flickr Commons


David Souter writes a weekly column for APC, looking at different aspects of the information society, development and rights. David’s pieces take a fresh look at many of the issues that concern APC and its members, with the aim of provoking discussion and debate. Issues covered include internet governance and sustainable development, human rights and the environment, policy, practice and the use of ICTs by individuals and communities. More about David Souter.