The irony was not lost on Twitter. Pictures circulated this week of shoppers queuing to enter the brand new “Amazon Go” store in Seattle, the whole purpose of which is to eliminate checkout lines.
The futuristic grocery store trialled by the web giant is packed with hundreds of cameras and tracking devices, so customers can saunter in while logged into an app, pick up the goods they want and simply walk out.
Payment for items is automatically deducted from their account upon exit — there is no need to deal with a human employee at all. Robotics, artificial intelligence and sophisticated video equipment are taking this Amazon store much closer to being worker‐free.
Inevitably, such a technological trial attracts interest. But the broader societal impacts concern policymakers. Amazon Go’s launch is really a microcosm of the debate about automation and its effect on labour markets. Last year a report by Cornerstone Capital estimated 7.5m retail jobs could be on the chopping board in the US alone due to these technologies, with 3.5m cashiers most directly affected.
Where retail leads, other sectors will surely follow. Pessimists project millions of jobs being rendered obsolete, in turn creating significant structural unemployment. Radical policy “solutions”, from having the government guarantee everyone a “basic income”, or even a job, have been suggested. At the very least, the consensus is the state should “do more” to prevent or alleviate the problem.
Technologies will no doubt transform workplaces over the coming decades, but commentators are surely over‐reacting. In technological terms, Amazon Go is really just the next long‐term leap from the widespread roll‐out of self‐checkouts and the proliferation of online shopping.
Consumers increasingly value convenience, and the supermarket sector is ultra‐competitive, pressuring retailers to cut costs. Amazon’s efforts are entrepreneurialism – testing whether there are yet enough customers out there who want to browse in‐store but who are significantly worried about wasting vast amounts of time in a queue to justify the investment.
The truth is that fears about automation have hung over the integration of machines in the workplace since the industrial revolution. The idea that “this time is different” due to the more sophisticated nature of machines is widely assumed. But so far these fears show up everywhere but in the data.
UK unemployment levels are at a four‐decade low of 4.3pc. The employment rate, at 75.3pc, is the highest recorded since it started being measured in 1971. US labour market participation rates have been scarred somewhat by the crisis, but there unemployment is low at 4.1pc too.
Maybe the pace of change will create big problems in future, but currently market economies with flexible labour markets appear adept at creating new opportunities as technology proliferates.
And this is unsurprising, for change is often gradual. When we hear stories about Amazon Go we jump to thinking about a state where all stores adopt the same technology. Reality is messier.
For now, the upfront cost of installing the technologies is prohibitively expensive for most retailers, and assistance from in‐house staff is still demanded by many. Change will be gradual rather than a “big bang”. Ironically, what really could accelerate this trend would be significant increases in the minimum wage (an implicit subsidy to automation) as demanded by Left‐wing groups.
None of this is to say that automation will not cause problems for some workers. All economic change, whether through altered trade or migration patterns, or in this case technological advances, will lead to direct job losses.
As the experience of mining in the UK showed through the Seventies and Eighties, some of those individuals affected will find it very difficult to secure new gainful employment, perhaps due to low levels of transferable skills.
But far too much is put on this downside and the “seen”, when the unseen benefits are much more significant. Let’s suppose that Amazon Go’s model becomes economic and is the default for the grocery sector. Yes, far fewer cashiers will be employed, and store assistants too will likely be replaced by machine technology.
But that does not mean there will be no employees. The development, maintenance and oversight of the new technology will require a smaller number of highly skilled staff, who will likely be much better paid than existing employees.
More importantly, the “unseen” effect is that the gains in efficiency in the grocery sector will filter through into the wider economy. Not having to designate significant amounts of stores to checkouts will enable more efficient use of space. Customers will benefit from saving time without queuing, which they can use to pursue other leisure activities or working. In the longer term, the labour‐saving nature of the technology is likely to mean reduced costs too, filtering through to lower prices than otherwise for consumers, given the competitive nature of the sector.
Such savings in time or money will be used by consumers for other demands, in turn creating new jobs in other sectors.
There’s good evidence already that occupations entailing personal human services (think personal trainers and social care) are in increasing demand, especially with an ageing population. But predicting exactly what people will do is to assume that labour markets can be planned or projected.
The question “where will the jobs come from” has echoed through time as new technologies have changed the nature of work and the structure of the economy.
But while recessions come and go, bringing with them temporary upheaval, there is little evidence that technology has ever significantly raised the unemployment rate, despite the widespread fear it will do so. While commentators fear the upheaval, the most likely outcome of new ventures like Amazon Go is higher standards of living, and changed patterns of demand.