Internet & Tech

Rise of the machines

How automation and artificial intelligence are reinventing the workplace


Wu Yulu and Number 25

WiC’s first references to China’s robot revolution featured people like Wu Yulu, a farmer-turned-inventor who merited mention in our fourth issue for his brood of machines.

“Each one is like my child so they all have my surname,” he explained to the local press. Number One goose-stepped around the village, Number Five smoked and Number Eight could do somersaults. Number 25, his favourite, pulled him around in a rickshaw. Those passed in the street were greeted politely: “Wu Yulu is my father. I’m taking him shopping. Thank you.”

Wu’s work was showcased in the press with other ‘farmer-inventors’ who had fashioned unlikely creations including homemade submarines and miniature aircraft carriers. But robotics today is a more serious proposition as China moves towards a future in which robots take more command.

Factories first

The robot revolution is already happening on production lines, where companies have started to hone down their human workforces. The attractions are alluring – no mistakes, no overtime pay and no calls for higher wages.

Research from Anderson Chow, Global Co-Head of Industrials Research at HSBC, has highlighted how payback periods on industrial automation have been shortening dramatically to not much more than six months. Prices are falling as robots are produced in larger numbers but the bigger factor is wages for factory workers, which continue to climb.

One company to watch is Foxconn, the contract manufacturer from Taiwan, which employs a million people in China assembling consumer electronics like the iPhone. It has been shifting some of its factories into lower-wage provinces but it has been trumpeting automation as well. Investors bought into the vision last summer, scrambling for the Shanghai-listed shares in Foxconn Industrial Internet, a subsidiary whose businesses include making robots for electronics assembly.

The rate at which Foxconn replaces its workers will tell a story about the robot revolution in general. It says it has already been able to lay off tens of thousands of staff, although it has missed some of its targets on automating a larger share of its workforce. Critics say that this is revealing and that the ‘Foxbots’ will take longer than planned to get the upper hand.

A buyer, not a seller

‘Robot density’ has been growing steadily in China in general: from 11 units per 10,000 employees 10 years ago to 97 last year, according to estimates from the International Federation of Robots, an industry group. This puts the Chinese in a similar bracket to the Americans and the Europeans, although some distance back on leaders like Singapore and South Korea.

China is already the world’s largest market for industrial automation, however, buying a little more than a third of all factory robots, more than Europe and the US combined. Car making and electronics, two sectors that account for about a third of purchases each in value terms, took the largest share, although ‘cobots’, or simpler robots that work alongside humans in warehouses and logistics roles, are getting more common too.

Expect sales to grow further and faster in China because so much of the world’s electronics manufacturing is rooted there and due to new demand from sectors like electric vehicles, which will trigger more automation.

The leaders in the heavy industrial space are still foreign-based firms, though, with Fanuc and Yaskawa from Japan, Swiss-Swedish giant ABB, and Kuka from Germany controlling about two-thirds of sales into the Chinese market. They are particularly dominant in the machines that do what is termed ‘3D’ work (too dull, dirty or dangerous for humans) like heavy lifting, gluing, welding and painting. Chinese companies are making better progress in smaller ‘delta robots’ for sectors like consumer electronics and communications equipment, plus the kind of cobots that work in logistics.

Domestic robot makers showed signs of losing share towards the end of last year, however, and weaker demand from the automobile and smartphone sectors has been putting pressure on the lower tier of homegrown firms. More broadly, HSBC’s Chow says that the domestic robot makers still need to master the manufacturing of key components, like the reducers that provide pinpoint accuracy for the most sophisticated machines (the Japanese have grabbed about three-quarters of the reducer business, he says). Two years ago Qu Daokai, president of Siasun Robot and Automation, one of the largest domestic robot makers by market value, said something similar when he described the domestic sector as “like a toddler” and complained about its reliance on foreign suppliers for key parts. Chow adds that the Chinese are yet to match ‘integrators’ like Fanuc in implementing automation across a complete production process too. “The Chinese are getting much better at making modular units for various points along the line,” he reports. “But they haven’t closed the gap on the integrators, who deploy system-wide solutions that incorporate third-party technology.”

Narrowing the knowledge gap

One strategy for catching up is buying the expertise in the way that Midea, from Shunde in Guangdong, has taken control of Kuka, a leading German firm, which was purchased for about $3.9 billion in 2016.

The takeover was controversial in Germany on concerns that trade secrets would be lost to a strategic competitor, although Chow counters that Kuka’s specialty in automation for car plants can’t be easily repurposed for factories that make washing machines and air-conditioners, where Midea is a dominant presence. Nor is it simply a case of absorbing the know-how of the foreign brands, much of which has taken years to establish. “It takes a long time to acquire the precision skills required in high-end robot production. Corporate culture is crucial and get-rich-quick plans don’t work, which is challenging for some of the entrants from China, who have shorter horizons for returns on investment,” he says.

Not that he is betting against the Chinese getting up to speed quickly: he believes that local robot makers will be posing much more of a challenge to their global rivals within five years.

Of course, there are examples of similar leaps: 20 years ago the Chinese were talking about railways that would revolutionise train travel; today their high-speed network is one of the most advanced in the world.

Another strong point for the Chinese is their focus on customisation, Chow says, citing the example of Shenzhen Innovance, an automation firm that has 40 sales teams dedicated to different types of industrial client. “Each of them has a technical head and a sales head, who meet clients together to define how best to meet their needs. This style of being ready to customise their approach to client needs is generally more flexible than the established players and the Chinese see it as a way towards making more sales,” he explains.

Facing the future in artificial intelligence

A related area where the Chinese have aspirations for leadership is artificial intelligence, where the focus is not so much on machines as machine learning.

On the factory line robots are designed to do something specific. In machine learning neural networks absorb information to make predictions with minimal human help. Deep-learning machines are capable of responding without human involvement, recognising patterns and taking decisions on their own.

So much for the theory but where is AI being applied most in real life? One of the main areas is facial recognition, where companies like SenseTime and Megvii are two of the unicorns in China’s AI herd.

Stories about the technology in action have featured high-profile cases like the dragnet for criminals at pop concerts, where cameras picked out the faces of fugitives from crowds of thousands. Customers at a KFC-owned restaurant in Hangzhou have been paying for their dinners with facial recognition, while a school even said it was monitoring its students’ expressions to determine whether they were paying attention in class.

Most deployments will be more mundane, including Beijing’s massive new airport, which is implementing facial recognition for its immigration queues and for matching baggage to passengers. Visual recognition will also be important in areas like cashier-less shopping, inventory checking, and quality inspections for goods.

Last year WiC talked to Matt Scott, the chief technology officer at Malong, another of China’s pioneering AI firms, about some of the other areas where breakthroughs are being made. Backed by Softbank, Malong is an international leader in image recognition and Scott picked healthcare as an example, explaining how Malong is helping doctors detect cases of ischemic strokes, where blockages of blood flows in the brain are immediately life threatening. “Doctors need time to review the CT scans and figure out what is going on,” he explained. “But our technology analyses the images instantly and flags the problem.”

Scoliosis, or curvature of the spine, is another condition where Malong is providing support. “Our technology looks at images of the spine and calculates what the best course of treatment should be. Of course, doctors can do this themselves but automating the process allows treatment decisions to be taken more efficiently,” he said.

Matt Scott at Malong: helping Shenzhen’s doctors

Information wars and arms races

One of the downsides to the robot revolution is the concern that it will put millions of people out of work. A counter argument is that machines will be working on tasks deemed too dangerous or too boring for humans (robot as a term was coined by a Czech playwright, from the word rabota, meaning servitude through forced labour). Additionally China’s population is aging and the once-limitless pool of workers that powered its economic model is drying up. Although some jobs will be lost, the defence is that many would have been surrendered to lower-wage economies anyway, often in sectors where China no longer wants to compete.

Instead the focus is more on hopes that productivity will pick up when humans work alongside machines. That could even result in higher wages for the ‘survivors’ in the workforce, a view partially supported by pay rises for some of Amazon’s warehouse staff last year.

Strategists from other countries worry more about China taking control of the robot world and especially that its machines might send sensitive information back to home base. There’s fear that the data that drives artificial intelligence could be abused as well, although the Big Brother way that the government is rolling out a ‘social credit’ system to rate its citizens’ behaviour has bothered overseas commentators more than the Chinese themselves.

Lee Kai-fu warns against an AI arms race

Another strand of the debate on AI sees it more as an arms race between China and the US, something that was discussed in Lee Kai-fu’s recently published: AI Superpowers: China, Silicon Valley and the New World Order.

Formerly a senior executive at Google, Microsoft and Apple, Lee is influential in Chinese tech and private equity. With a foot in both camps of what he describes as “the great AI duopoly”, he has an unusual ability to talk about their respective strengths and weaknesses.

He draws an interesting distinction between ideas and implementation in the sector, putting the Americans ahead in the pure science behind AI, but positioning the Chinese as the pacesetters in how the theory is being applied.

In particular, he points to the massive potential for Big Data in China, much of it derived from the world of social media, like WeChat, the country’s killer app.

His premise is that the bigger and more diverse the dataset fed into the algorithms, the better they will fine-tune their performance. That is going to eat into the American lead in the science of AI. “Having a monopoly on the best and the brightest just isn’t what it used to be,” he believes.

Asked about this ‘arms race’, Matt Scott from Malong pushes back against the idea of two separate worlds, arguing that the mood in the research lab is much more collaborative. He has a similar view about his clients, saying that they don’t think in national terms, only how AI will help their businesses.

“In sectors like healthcare the benefits will extend across borders, including the imaging projects we are running with hospitals in Shenzhen. The work there is going to develop models that help millions of patients around the world, not just in China,” he adds.

Lee still comes back to the same story of superpower rivalry in much of his book, talking about areas in which each side has an advantage. The Americans have a commanding lead in the technology behind ‘autonomous AI’, for example, which will give them an initial edge in new applications like self-driving cars. Support from the state in must-win battlegrounds like cloud computing, Big Data and 5G technology could give Chinese firms the means to catch up, however.

Lee also predicts a growing sense of separation between the two countries that makes it challenging for artificial intelligence firms to prosper in both. “It is very difficult, if not impossible, for any American company to try to enter China’s market or vice versa,” he explained to reporters. “It’s like two different jigsaw puzzles. You can’t take a piece from one and try to fit it into the other – everything is different.”

© ChinTell Ltd. All rights reserved.

Sponsored by HSBC.

The Week in China website and the weekly magazine publications are owned and maintained by ChinTell Limited, Hong Kong. Neither HSBC nor any member of the HSBC group of companies ("HSBC") endorses the contents and/or is involved in selecting, creating or editing the contents of the Week in China website or the Week in China magazine. The views expressed in these publications are solely the views of ChinTell Limited and do not necessarily reflect the views or investment ideas of HSBC. No responsibility will therefore be assumed by HSBC for the contents of these publications or for the errors or omissions therein.