Over the last few months, we have come to depend on the UK’s ‘Key Workers’. Their German equivalents are the ‘systemrelevante Arbeitskräfte’; a term which, while lengthy, somehow bestows greater respect on the individuals who have kept us fed, safe and clean during lockdown.
Among already well-respected professionals such as doctors, nurses, the emergency services and members of the armed forces, key workers also include those who may have been invisible to us as we formerly went about our daily business. During lockdown, however, we noticed anew as they delivered our letters, removed our rubbish and stacked supermarket shelves. They pulled round the clock shifts in warehouses to fulfil our one-click whims and then delivered the sea of brown boxes created as a result.
Judging from the thank you notes pinned to front doors, we now value these workers more highly. Hopefully, this newfound respect will last. But applause can’t mask the fact that many of these individuals aren’t paid a living wage, are on zero-hour contracts, or are part of our gig economy. Many are included in the 4.2 million people in the UK on ‘low pay’, and who make less than two-thirds of the median pay of £13.21 an hour.
And while some kept their jobs during the recent pandemic, the lowest earners amongst this group were more likely to have lost work or been furloughed due to the crisis. Those who did remain in work were put at a higher risk of infection as their jobs invariably placed them at the frontline of the fight against COVID-19.
In some respects, then, we should welcome the arrival of medical robots that were unveiled recently in Rwanda. The shock troops of a new wave of automation estimated to be worth $8 trillion over the next 10 to 20 years in the US alone, these UN-supplied ‘anti-epidemic’ robots are being put to use in the fight against COVID-19. They can test the temperatures of patients and detect those who are not wearing a face mask, removing the need for humans to put themselves at risk to undertake these tasks.
In appraising these robots, look past the rushed industrial design suggestive of a battle between product development and marketing over whether the robot should resemble a human or a fridge. Ignore the fact that the result looks like an iPad balanced precariously on a swing bin full of faulty, blue LED Christmas lights. Instead, consider that the robot can “monitor abnormalities in how patients sound or look.” Nothing sinister there, then.
Even if you aren’t concerned about the integrity and transparency of the algorithms that form the basis of their clinical decisions, give a thought to the inherent bias that underpins much of the machine learning on which these robots’ ‘artificial intelligence’ (AI) no doubt depends. Is it the by-product of inevitably limited data sets and the programming skills of young men with bad diets in Iron Maiden t-shirts, and all that entails? Maybe. Regardless, remember that while AI can beat a human being at the ancient game of strategy, Go, it is also the reason that millions curse Siri’s very existence every day for failing to answer a simple question about the opening times of a local supermarket.
Indeed, the sheer difficulty of automating the many tasks humans take for granted is why “away from the research labs, expectations around driverless cars are cooling”, for example.  AI lacks the human ability to use “top-down reasoning” to make sense of confusing or contradictory situations; poorly marked road signs, snowy roads and inconsistent curb heights can cause the AI in a driverless car to freak out. Imagine what it does when faced with the human body in all its glorious complexity.
Still, the robots are coming, and we had better get used to them. Bain & Company warns of a volatile decade ahead, created by a toxic combination of demographic shifts, automation and inequality. This could see between 20 to 25 per cent of jobs eliminated, with the greatest impact being felt by — you guessed it — middle to low-income earners.
The productivity gains from automation are simply too significant to ignore: China’s largest employer, Foxconn, eradicated 60,000 jobs, or 55 per cent of its total workforce, in one of its factories alone in 2019. Low-level clerical jobs are apparently simple to automate.
So-called ‘cobots’, which augment humans in the workplace, are designed to support workers by handling tasks that demand greater strength, agility, stamina or focus for repetitive tasks. And while they don’t necessarily replace humans entirely, the cost of this technology has fallen so dramatically that the average hourly cost of a cobot is 12 times cheaper than a German manufacturing worker. It’s reaching the point where cobots are nearly as cheap as manufacturing workers in China or India.
AI is even impacting traditional white-collar jobs such as legal services. The pace of change is staggering, and it is expected that workforce changes in the 2020s will be on a par with those created when agriculture was automated between 1900 and 1940. Although that took forty years, not a decade, as current estimates have it.
The essential workers we have come to rely on during the pandemic — from delivery drivers to nurses — are at considerable risk of being replaced by machines, whether they are parcel-bearing drones or passive-aggressive swing bins.
Around the world, massive fiscal and monetary stimulus helped save jobs during the COVID-19 pandemic. But as furlough comes to an end, and businesses strip costs to meet limited consumer demand, the outlook for jobs doesn’t look great. Unless we all gather round a Modern Monetary Theory debt bonfire, the massive bills racked up by Governments during lockdown will need to be paid off, and that will be hard to do in economies where productivity has stalled.
Automation will, of course, greatly improve productivity, but at a high price. US automotive manufacturing productivity grew by 128 per cent between 1993 and 2014 as greater automation took hold, yet employment in the sector fell by 28% over the same period. Given that populism was in part a response to fears of job losses from increasing automation, what Government would now be seen to favour robots over voters?
As others have said, we can’t uninvent technology. And while the air is heavy at the moment with calls to use the pandemic as a catalyst for social, economic and environmental change, gritty reality may trump these grand visions. I can’t see a Universal Income to counter the effects of demographic shifts that will favour saving rather than spending becoming any more likely, for example. What chance more carbon taxation to fund the stimulus? Not much, I would have thought.
I’m afraid the robots, the productivity gains they bring, and the job losses they will herald, are on their way. We may not want or necessarily need them, and we may be right to fear for our jobs and feel frustrated at the lack of debate around their arrival. I can see some benefits to automation. Although I can’t imagine developing a relationship with a droid in the same way that we have with our Amazon delivery driver, who knows our first names and offers a friendly smile when we see him each day.
No, the robots are coming. Just don’t expect me to go out on the street and clap for NHS droids every Thursday evening.
 The effects of the coronavirus crisis on workers, Resolution Foundation, 16 May 2020, https://www.resolutionfoundation.org/publications/the-effects-of-the-coronavirus-crisis-on-workers/
 The five: robots helping to tackle coronavirus, The Guardian, 31 May 2020, https://www.theguardian.com/technology/2020/may/31/the-five-robots-helping-to-tackle-coronavirus
 Driverless cars show the limits of today’s AI, The Economist, 11 June 2020 https://www.economist.com/technology-quarterly/2020/06/11/driverless-cars-show-the-limits-of-todays-ai
 Labor 2030: The collision of demographics, automation and inequality, Karen Harris, Bain & Company, https://www.bain.com/contentassets/fa89826544934e429f7b6441d6a5c542/bain_report_labor_2030.pdf