Pasar al contenido principal

Last week I wrote about some of the policy and regulatory issues that arise from the accelerating trend towards a digital society and new technologies such as artificial intelligence.

This week some thoughts about employment issues that arise from that same trend: issues that are increasingly being discussed but on which we have very little evidence to base conclusions.

Visions of the future

As often, visions of the future are divided between optimists and pessimists. Where employment is concerned, the former believe our children may enjoy boundless leisure while robots do what once were jobs around them. Pessimists, on the other hand, expect robot-induced unemployment to leave them impoverished and losing their sense of worth.  

As usual, the future's likely to lie somewhere in between. I’m not going to suggest exactly where but will make six points that strike me about employment in the coming digital society.  My thanks to colleagues at a recent workshop for inspiring these.

How many workers?

First, what are the employment impacts of AI that are feared today? Let’s start with numbers.  

Many different estimates have been made about potential job losses (and even, sometimes, increases) arising from AI. Many are made by those with vested interests, not least by consultancies whose future contracts depend on hope, which often leads to hype. Others, often made by academics, are more gloomy. A key issue is whether analysis should focus on whether it's whole jobs that may be taken by technology or rather tasks within those jobs, only some of which may be susceptible.

Myself, I think there’ll be job losses, which may be large, but none of the current estimates convinces me; they’re far too speculative. Policymakers, in my view, shouldn’t bet the future on them. They should monitor the impact of new technology on jobs with care, precision and granularity – looking at different places, different skill levels and social groups. They should commission independent foresight analysis. And they should prepare for different outcomes, not only 'good' or 'bad' ones.

What kind of jobs?

Employment isn’t all to do with numbers. Job quality matters too. The casualisation of labour (which some call flexibility) makes jobs more insecure. Workers in the so-called gig economy don’t share the employment rights and fringe benefits (such as pensions and maternity leave) that have been won by workers through generations of campaigning, solidarity and social legislation. Workers in AI-enabled roles don’t look as likely as traditional employees to have them either.

These aspects of employment don’t add value just for individuals; they’re important for society as well. The more people lack the social protections that traditional employment offers, the more society in general has to provide a safety net (if it can afford to do so). This is particularly important as AI seems likely to polarise jobs, hollowing out those with middle-level skills and dividing workforces more sharply between high and low earners.

We need to rethink how employment fits with social welfare in the coming digital age.

Is there an international framework for employment rights?

Yes, in one sense. Employment rights make up a large part of the International Covenant on Economic, Social and Cultural Rights, a core part of the international rights regime. But the framework it contains, based on assumptions about formal employment and the presence of trades unions, was based on European and North American experience. It looks out of date and out of kilter now.

Most workers don’t, today, have jobs for life with one employer, even in the North (and it’s doubtful that they ever did). Most are no longer union members. Most people’s work in many countries is in the informal sector, ungoverned and unregulated.

We’re going to need to rethink employment rights to address the informality and flexibility of the digital age. This requires governments, employers and unions to change their ways. Unions in particular, if they're to have much future, need to find new ways of representing workers who don’t have traditional working lives.

What will happen to the fruits of job displacement?

Even those that think there’ll be more jobs arising from AI recognise that old jobs will go and this will be disruptive.

The argument that’s often made by AI advocates is that machines and algorithms will take over the boring tasks, which they can do more efficiently than people, allowing human workers to be redeployed to more creative, imaginative and fulfilling roles.

That may happen in businesses where there’s scope to develop new markets and new services. But many businesses and organisations are likely to bank efficiency savings by cutting labour costs (and jobs).  This seems particularly likely in the public sector because of the demand from voters for politicians to cut taxes. It’s this, I think, that makes unemployment seem more likely to result than redeployment.  

That in turn could have adverse effects on aggregate demand in the economy: if fewer people are employed, then fewer people can afford to buy the goods and services that AI may generate.

Is reskilling folk the answer?

It’s generally agreed that reskilling – and continuous reskilling throughout lifetimes – will be essential for the future workforce. A lot of governments have made policy statements about building skills for the digital age, often focusing on school curricula, on higher education and on lifelong learning.  

But here’s an irony. Analysts of AI and jobs also suggest that the loss of jobs/tasks that machines can do better than us (especially in manufacturing and services) will lead to more jobs/tasks being created that we can do better than machines (like caring and creating).  Three comments on this:

One: it’s unclear where the funding’s going to come from, especially for lifelong learning. Will this be a responsibility of government, business or the individual?

Two: a lot of emphasis is put on teaching coding skills in schools. I’ve no objection, but I doubt that many future workers will use coding in their working lives (any more than most of us use algebra today). Jobs that use IT to do things don’t necessarily require technical skills (and that’s likely to become more so as algorithms take more responsibility for writing algorithms). Information literacy, critical and creative thinking strike me as more important.

Three: I’m sceptical about the scope for growth in caring and creative work (both of which are, incidentally, in general low paid). And, surely, if it is expected that these are going to be the growth areas for employment, government strategies for skills development in the digital age ought to be prioritising them rather than (or at least in addition to) more digital-specific skills. I know of no strategy for skills development that does so.

What of a post-work world?

Lastly, what if we’re heading for a post-work world, with fewer jobs and shorter working weeks for all?  There’s quite a lot of interest expressed now in new forms of income distribution, such as giving everyone a universal basic income whether they’re in work or not.  I personally can’t see that gaining political resonance in the North or being financially feasible elsewhere. Nevertheless, policymakers need to consider options for maintaining incomes at a time of radically different, potentially reduced and more unequal distributions of work than we’re accustomed to today.  

This isn’t just about the individual

Too much discussion of the Information Society focuses on the individual rather than society. Impacts on employment will have lasting legacies on communities as well as individuals.  Workplaces are social environments as well as income sources. Their loss undermines social coherence as well as individual and household welfare.

Even if a country as a whole negotiates its way through large-scale job changes such as may be coming soon, some regions, towns and communities are likely to be left behind.  I grew up near to a town called Sunderland in the north of England. The three main industries that supported it and local villages – coal, steel and shipbuilding – disappeared in the 1980s. Those communities have not recovered fully yet from the loss of jobs and, as important, self-confidence. 

What about the workers?

One final point. Too much discussion about jobs and AI takes place among academics and policymakers who have little or no experience of the jobs which are likely to be most affected – middle-ranking professional and skilled and unskilled manual and service jobs.  

I heard a leading development economist say recently that everything would be fine because AI would just get rid of boring work which no-one wants to do.

We’ll get nowhere if policymakers are so patronising. Manual and service workers are no less skilled than development economists. They just have different skills and got paid less for doing something that is often much more dangerous. Most people who do the routine jobs that this development economist despised gain satisfaction from them, feel valued in them, and would be devastated by their loss. And most people would rather have a boring job than have no job at all.  

Experience in the workplace, like most things, can be both good and bad. Policymakers should listen to the workers when they’re trying to decide about their futures.  Better still, involve them in deciding.

Inside the Information Society will be taking its annual Northern winter / Southern summer break, and will return in January. 

Image: By Daniel Wiadro on

Read also: 2019 Global Information Society Watch on Artificial Intelligence, human rights, social justice and development 
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.