a robot

Robots were used by Alibaba to run its supermarkets but with mixed results

Imagine walking into a supermarket with no staff. You order products using QR codes, pay using Apple wallet, watch a robotic arm deliver your items to you, then simply ‘grab and go’. 

It may sound like the stuff of science fiction, but this revolutionary approach to retail was actually pioneered by the Chinese tech giant Alibaba back in 2017. 

The company hoped that by reducing the cost of human labour, the plan would save both time and money. Commentators also predicted that it could make millions of Chinese workers redundant. 

As a worker in Hangzhou, ‘the Silicon Valley of China’, I watched with interest. 

Would these unmanned supermarkets change the retail landscape in China and beyond? And if so, what would become of the millions of shop workers whose jobs were replaced by robots? 

When I returned three years later, things had not progressed as many predicted. Alibaba’s supermarkets had been a significant success, with the firm establishing itself as one of the key players in Hangzhou’s grocery sector. 

However, these supermarkets were no longer devoid of human helpers. On the contrary, they were now heavily staffed. Store assistants were on hand to welcome shoppers, give out samples, stack and tidy shelves, and help customers pay. 

While all of these takes could be performed by intelligent machines, Alibaba had found that human workers were more suitable for this type of business. 

One reason was that most people may prefer to be served by fellow human beings. Older customers, in particular, were struggling with the new technology and choosing to shop elsewhere. 

Others were deterred by concerns about what would happen to their shopping – and their money – if they needed help, or if a machine broke down and there was no-one around to assist them. 

It seems that public-facing jobs, where customers wanted the experience of being served and cared for, and that required a greater degree of flexibility, needed to be done by humans. However, another factor was the cost. 

Contrary to expectations, replacing human workers with robots did not produce significant savings in time or money. The cost of setting up these machines was a huge investment. 

So was maintaining them. When a robotic arm broke down, it took two hours to send someone to investigate the issue. Then more hours to send the parts for a replacement. 

The problem also had to be fixed overnight, when there were no customers around, to maintain the flawless, futuristic image of the stores and the technology they used. Otherwise the benefits of the new technology would be less clear and attractive from a customer perspective.

This increased labour costs and caused supply chain issues if further parts were required when the rest of the business was closed. 

By comparison, replacing a human worker was simple. When one rang in sick, managers simply called in another to cover their shift. Problem solved. 

The disadvantages of robots vs humans

Two firms that attempted a similar approach collapsed before the pandemic. It is likely that Alibaba lost money and investment at a similar rate to its competitors, but it had the capital to survive and adapt to a more sustainable model. 

Does Alibaba’s retreat from the idea of ‘unmanned supermarkets’ spell the end for robots in retail? Not at all. Many of the changes attempted by Alibaba are mirrored at other firms who are exploring AI as a way to cut costs and gain a competitive edge over their rivals. 

Take Amazon, for example, which occupies a similar position in the West as Alibaba does in China. Robots and intelligent machines are becoming more common across its operations. 

Has this made Amazon workers more vulnerable about their jobs? Again, no, at least not excessively so. Conducting interviews with warehouse workers and their managers in recent months, it was surprising to see the diverse range of opinions on offer. 

While some had concerns, most frontline workers that I spoke to believed these digital technologies had actually created jobs and improved some aspects of their work, with some highlighting how assistance from robots had made their labour less physically demanding. 

Some staff had even joined from other warehouses, despite the rise in robotic technology, attracted by Amazon’s better pay. 

Managers revealed another side to the story. Implementing robotics had slashed the workforce by as much as 90 per cent at individual sites – at one warehouse 60 workers were reduced to just six.  

But as Amazon’s business expanded and warehouses continued to be established, a recruitment drive had camouflaged those redundancies. Amazon UK created 25,000 permanent jobs across corporate functions, R&D, and operations in 2021. That number was expected to grow last year. 

In addition to this, some Amazon workers were being upskilled to supervise the newcomers (both robotic and human). For the time being at least, these robots appear to cost-saving job producers.  

Technological innovations might automate some tasks, but do not necessarily replace human workers. 

In China, an abundant and cost-effective workforce has given human labour an advantage over Alibaba’s robots, which proved to be expensive and time-consuming to install and maintain.  

Meanwhile, Amazon’s continuous growth in the UK has outweighed the diminishing demand for staff in the labour process at each individual workplace. As labour costs rise, intelligent machines could become more attractive, supplementing and even helping to grow the UK workforce. 

In both cases, job creation has grown alongside the increased automation. 

How will robots change work in the future?

So what does this mean for workers on the frontline? I believe that it means we have good reason to be hopeful. 

My findings suggest that “technological unemployment” – job loss directly caused by new technology – is not inevitable. 

Some commentators continue to make predictions about the future of work based on whether individual tasks could be automated in a particular profession. I would argue that this, by itself, is not a good indicator for the real risk of job loss and often predicts the ‘worst case scenario’. 

In order to properly understand the impact of automation and AI on employment, it is essential to consider the political economy of technology in a particular workplace. 

On the other hand, staying hopeful is not enough either. Only action will provide workers with the security they need. 

Giving workers a voice about individual experiences gains greater importance in this new, AI-enabled world of work, because it helps to reveal the actual impact of technological transformation. This might be achieved individually through social media platforms, or collectively through union representatives. 

What we need is more empirical research. That is why WBS has teamed up with the Institute for the Future Work to collect first-hand data to understand how technologies impact the UK job market. Our goal is to provide robust evidence to guide Government policy.  

After all, this is one of the decisive factors that will shape the business environment and labour standards of the future. This, in turn, will determine how technology is negotiated in the workplace while protecting the jobs and livelihoods of the human workforce.

Further reading:

Liu, H. Y. 2023. The role of the state in influencing work conditions in China's internet industry : policy, evidence, and implications for industrial relations
Journal of Industrial Relations, 65, 1, 3-21.

Liu, H. Y. 2023. Digital Taylorism in China’s e-commerce industry : a case study of internet professionals. Economic and industrial democracy, 44, 1, 262-279.


Hong Yu Liu is a Postdoctoral Research Fellow at Warwick Business School and holds a Research Fellowship at the Institute for the Future Work. He teaches the Entrepreneurial Mindset on the Undergraduate programme.

For more articles on Finance and Markets sign up to the Core Insights newsletter here.