Canaries in the coal mine

Today we’re launching an interactive heat map presenting the results of our research which predicted the impact of AI-fuelled automation in individual parliamentary constituencies in England, Scotland, and Wales.

Jobs at high risk of automation in Great Britain – heat map

Since we first released our findings in our report in October last year, there’s been a bit of push-back against these kinds of predictions that there will be high levels of job losses from AI-fuelled automation over the next 15 years.  

The main thrust of the push back is that automation will happen more slowly (or not at all) because once you break jobs down into their component tasks there remain many things that humans will be better at (or cheaper at) than machines for a long time. So apparently we need to relax a bit.  And anyway, lots of new jobs are going to be created. Everything will be OK with a bit of retraining and education. This certainly seems to be the dominant narrative in Westminster now, with the Prime Minister, Lords AI Select Committee, and others all striking a relatively optimistic note on AI and jobs.

But developments in the UK this year suggest that things may be moving faster than even we predicted. Our report, which built on cutting edge modelling by PwC and others, predicted that automation would be particularly rapid in certain sectors including retail and call centres.  Recent news headlines would suggest we were right. Job losses in these sectors should be seen as canaries in the coal mine.  Early indicators of explosive change to come.

Retail, warehousing, call centres, and other canaries

Politicians need to start joining the dots.  It seems not a day goes by without the announcement of store closures and job losses by one of the titans of the British high street. Marks and Spencers, Mothercare, Carphone Warehouse, House of Fraser, Poundworld:  the list just keeps growing.   What’s that got to do with automation? Everything.  The client-facing part of retail is increasingly automated and delivered online, rather than in store by people.  

Meanwhile, more and more of retail’s back end is also being automated.  Videos of Amazon’s automated warehouses have a balletic beauty.  But they represent an ugly truth for workers in warehouses across the UK.  Amazon’s competitors will have to automate or die. And that means more job losses.  In April Shop Direct announced the closure of 3 warehouses in Greater Manchester which employ 2000 people.  The warehouses will be replaced by one automated warehouse in the East Midlands where about 500 new jobs will be created.  Sainsbury’s and Asda’s merger can also be put down to the Amazon effect. Asda’s parent company Walmart is retreating to focus on fighting Amazon at home in the USA.  In the UK, the companies claim there will be no closure of stores. But is seems certain that there will be a great deal of consolidation and automation of wider operations, including back-office processes.

Change is also coming in the call centres which, over the last decades, have picked up the employment slack created by de-industrialisation in Britain’s heartlands.  There are 6,200 ‘customer service centres’ in the UK employing nearly 1.3 million people. Call centres in the South are prohibitively expensive to run. But large talent pools, reassuring accents (pardon the cliche), and low costs led to a boom in Wales and the North.  

But all this looks set to change dramatically as automated systems are able to deliver an increasing proportion of call centre tasks.  The closure of Tesco’s Cardiff call centre, eliminating 1100 jobs in February was quickly followed by Virgin Media’s announcement that they would close their call centre in Swansea eliminating 800 more.

There are some themes that link all these recent job loss stories.  They are jobs that involve repetitive tasks in relatively predictable environments.  They are in highly competitive cut-throat sectors of the economy with narrow profit margins where companies cannot afford to let their competitors get ahead of them in terms of costs.  

Automation optimists rightly point out that millions of jobs will be created in the future, jobs doing things that we can’t even dream of now (how many people would have understood the job title “social media officer” 15 years ago?) and so we shouldn’t worry too much.  We just need to make sure people get the right education and re-training to prepare them for the future. And it is certainly true the employment rates are at a record high (although the big growth has been in precarious work in the gig economy).

But it seems to me that the pace of change is going to be ferocious.  It is going to hit certain people in certain places very hard, and these places are not necessarily where the new jobs are going to be created.  The areas we identify as being at high risk of job displacement have suffered previous deindustrialisation, with the collapse of, amongst others, the coal mining industry. The socioeconomic impacts of these industrial shifts are still being felt today, with many of these areas suffering high unemployment and poor health outcomes.

Political rewards

Politicians must not bury their heads in the sand.  They need to develop smart, targeted strategies to address future job displacement and support those people and places that are being hardest hit both financially and psychologically.  In the longer term we will need new models of taxation and redistribution and a new attitude towards work in a world where there may be less work to do as machines become better than humans at more and more intellectual tasks as well as physical tasks.  This need not be a social disaster, the future could be bright, but it will require real effort and ingenuity to get us there. There will be great political reward for the party that gets this right.

The question we should be asking ourselves is not ‘how is the AI revolution different from previous industrial revolutions?’ but rather ‘how do we make sure that the AI revolution is different from previous industrial revolutions?”


Olly Buston