Sunday, May 05, 2019
Robots aren't destroying jobs, but nor are they boosting productivity
Posted by David Smith at 09:00 AM
Category: David Smith's other articles

My regular column is available to subscribers on This is an excerpt.

The future is, by its nature, something that is yet to happen. In some ways though, what we think of as the future – things like artificial intelligence (AI) and robots – are already with us. So, without wanting to sound like some terrible advertising slogan, the future is now, and that allows us to make an early assessment of its impact.

To pessimists, new technologies like AI and robotics are always going to be a destroyer of jobs on a vast scale. People will be replaced by machines that, in many cases, will be a lot smarter than they are.

I have always argued, by the way, that while some jobs might be replaced by machines employment would not be. Technology would in future create as many jobs as it destroyed, if not more. It is not the best example but the cottage industry that has grown up to replace smashed smartphone and tablet screens, and can be seen in every high street and shopping centre around the country, clearly would not exist if there were no smartphones and tablets.

The optimistic argument about AI and robotics goes a lot further than “they won’t bring mass unemployment”. The argument here is that these technologies will unleash a period of strongly rising productivity and prosperity, shaking us out of our post-crisis lethargy and into a much brighter and better future. The march of the machines, in other words, should be something we welcome with open arms.

As it happens, a couple of new reports, one from the Organisation for Economic Co-operation and Development (OECD), its latest employment outlook, and the other from the Chartered Institute of Personnel and Development (CIPD) speak to this very subject.

The OECD’s outlook is called The Future of Work and goes into every aspect of the labour market effects of new technologies. It notes that we are in an era of more rapid diffusion of technology. In America it took seven decades for the proportion of households with landline phones to rise from 10% to 90%, while for mobile phones it took only 15 years, and for smartphones eight years. Business spending globally on information and communication technologies (ICT) has risen rapidly and there has been a fivefold increase in sales of industrial robots this century.

Across the 34 members of the OECD, 14% of jobs are said to be at high risk of automation and 31.6% at risk of significant change as a result of new technologies. For the UK the figures are 11.7% and 26% respectively. “The manufacturing sector is at high risk, but so are many service sectors.” It says. “And, even though the risk of automation is low in health, education and the public sector – many people will be affected because those sectors employ a large share of the workforce.”

On automation, the OECD agrees with me that we should not confuse the risk to individual jobs with the risk to overall employment. As it puts it: “Despite widespread anxiety about job destruction driven by technological change and globalisation, a sharp decline in overall employment is unlikely. While certain jobs may disappear, others will emerge, and employment has been growing overall.”

This does not mean we can relax entirely. People will need help, and training, to transition out of the jobs that are being replaced by machines and into new ones. Some workers, and some regions, are at greater risk. Young people, particularly those with no post-school education, are at the greatest risk says the OECD.

The CIPD, in a survey carried out with PA Consulting, found that a third of 759 businesses it questioned had invested in AI and automation over the past five years. The employment effects were, contrary to fears, positive; 35% saw more jobs in the areas affected, and 25% less. By nearly three to one, 44% to 18%, employers said AI and automation had made jobs more secure. Among employees, more than half, 54% said these new technologies had not helped them do their job better. There was little evidence that AI and automation were associated with significant productivity improvements, which the CIPD puts down to “lack of thought and planning on how people and technology are working together”.

It is, as I say, early days but it is fair to draw some initial conclusions. There are sector like retailing, which saw a 2.4% drop in employment in the year to the first quarter, where an effect on jobs from automation, such as customer-operated checkouts, can be detected, though separating that from other factors affecting jobs is difficult. In Northern Ireland I came across a firm that had no alternative but to automate because of the loss of EU migrant workers. I am sure there are many others.

Overall, however, it is hard to detect an employment effect. Britain’s employment rate is at a record high. At 76.1% it is near the top of the OECD league table, behind only Germany, New Zealand, Japan, the Netherlands, Sweden, Switzerland and Iceland. But, in Britain and elsewhere, there is a stubbornly high proportion of people low-skilled jobs. Automation, which might be expected to render many of these low-skilled jobs redundant is going hand in hand with an increase in their number. To be optimistic about the productivity consequences of new technology, you would want to see many of these low-skilled jobs automated out of existence.

The OECD has some other interesting comparisons. The UK has a lower proportion of temporary workers than average but a higher proportion of part-time and self-employed workers. Flexibility operates in many different ways.

What it is not doing, so far at least, is producing any evidence at a macro level that AI and robotics are bringing an increase in productivity. Though the picture in the UK is worse than most for productivity, it is hard as yet to detect any sign that the optimistic technology story, the one in which it brings noticeably stronger increases in productivity and prosperity, is happening anywhere in the OECD. The optimistic technology story, in other words, has yet to reveal itself.

That may be because we are still in the foothills of shifting the UK and other western economies into a higher gear. I am sure there are consultants reading this who can point to discernible, technology-driven productivity gains for their clients. It may be that the pioneers of AI and robotics are by trial and error merely paving the way for others who will be able to drive meaningful and lasting productivity gains and a future of rising prosperity, more leisure time and more people in skilled rather than unskilled jobs. That has always been the great promise. We can only hope it becomes a reality.