// Report on the Assemblages of Humans and Machines in the Pearl River Delta
By Merve Bedir, Marten Kuijpers and Marina Otero Verzier
This article is part of “Automated Landscapes,” a long-term collaborative research initiative on the implications of automation for the built environment, launched in 2017 by Het Nieuwe Instituut, and directed by its Research Department. The project addresses the contemporary emergence of distinct types of spatial configurations and conditions engendered and afforded by automation, focusing on present-day case studies in the Netherlands and in the Pearl River Delta region.
In 2011, Foxconn—the world’s largest electronics contract manufacturing company, best known for producing iPhones—announced a plan to equip its factories with one million robots by 2014. As CEO Terry Gou has noted, replacing human labour with automation technologies for simple, routine tasks would decrease costs and improve efficiency. (1) Like Foxconn, an increasing number of companies in China argue for the introduction of robots to their production lines. Yet in 2016, there was an average of only sixty-eight industrial robots per 10,000 workers in China, in comparison to 631 in South Korea and 309 in Germany. While China still lags behind countries such as Korea, Germany, Japan, and the United States, its development and implementation of labour automation technologies is rapidly increasing. (2) Since 2013, China has been the biggest market for robots in the world.
At the epicentre of this transformation is the Pearl River Delta region, which, as home to companies such as Foxconn as well as tech giants including Huawei and ZTE telecommunications and BYD electric vehicles, is currently the world’s dominant human-powered factory. The region is becoming a key arena for the transition towards manufacturing systems based on human and robot interaction. From automated kitchens and tailor shops, to robotic arms and agricultural drones, the spatial arrangements and protocols that result from labour automation challenge conventional spatial requirements and regulations for health, safety, and welfare. These automated spaces serve as flagship projects for the region’s larger ambition to position itself as a hub for innovation and sustainable smart technologies. Yet, in addition to increasing the efficiency of the factory, robotics brings new forms of labour ethics and structures and territorial occupation that impact the city, the Pearl River Delta region, and ultimately the bodies that inhabit them.
Support for the transition to automation in China, from both companies and the government, was triggered by the country’s economic slowdown, which was directly attributed to its manufacturing industry. The existing export-oriented production model based on an infinite supply of cheap labour became a hamper, particularly in a context of changing workforce dynamics and rising wages. (3) Automation was thus regarded as an alluring alternative. In 2014, the Chinese government launched Made in China 2025, a national, ten-year plan to shift the driving focus of its manufacturing industry from a paradigm of quantity to one of quality. The plan envisions an increase in labour productivity by means of revamping factories with advanced technologies such as robotics and smart automation. (4)
Similarly, in the past few years, local and regional governments have initiated their own program for transitioning to automation, and regional governments around the Pearl River Delta are at the forefront of this development. (5) In 2015, Guangdong province announced a subsidy program with the revealing title “Robots replacing human workers.” The initiative focuses primarily on the adoption of automation in companies involved in the production of automobiles, motorcycles, home appliances, electronics, construction materials, and clothing. (6) Between 2015 and 2018, 943 billion Chinese Yuan were allocated to this program, with subsidies covering ten to twenty percent of a company's industrial robotics upgrades. In addition, city governments in Guangdong province, such as Dongguan, Foshan, and Shenzhen have initiated pilot projects.
To position China at the forefront of the so-called fourth industrial revolution, Made in China 2025 focuses on upgrading the industrial manufacturing sector from the “Industry 2.0” model, based on mass production using assembly lines, to the “Industry 3.0” model, organized around smart manufacturing. Whereas countries like Germany and the United States see this new industrial revolution as an opportunity to bring manufacturing back within their borders, China sees it as a strategy for developing its industrial sector into one which will remain competitive in the long run. Furthermore, while a growing number of new, fully-automated factories are launched in Western Europe and North America, in China automation technologies are gradually being introduced into existing human-powered workplaces.
However, the challenges of integrating automation become clear in the case of Foxconn factories. Of the one million robots that Terry Gou proposed to install, only 50,000 were implemented three years later. According to the general manager of Foxconn’s automation technology development committee, Day Chia-Peng, the endeavour proved more arduous than expected: aside from the difficulties of replicating human hand-eye coordination, any robotic solution needed to be constantly adjusted as product cycles in the electronics industry continue to decrease. (7) Following this experience, Day strongly advised local governments not to regard robots as the ultimate panacea for labour shortages. (8
Furthermore, while the Made in China 2025 program is explicit about its potential economic impact, it does not project how the rapid introduction of robots and artificial intelligence (AI) might affect China’s labour market and income distribution. (9) Organizations such as the Ministry of Labour and Social Security, the Ministry of Education, and trade unions were not consulted in the development of the program. (10) Yet, recent studies on current automation projects and related policies indicate that this transition to advanced manufacturing is likely to result in massive layoffs and profound societal transformations. (11) Some of the displaced factory workers might be able to find work in other sectors or to work their way up in the factory, but these options won’t be within reach of all workers, many of whom are migrants. While a variety of measures have been proposed to compensate the consequences of increased automation—including educational programs to develop workers’ skills and competences, a universal basic income, and alternative tax systems—the main focus of Made in China 2025 is on technology.
Research about automation's influence on labour markets is widely available, while its impact on the configuration, design, and occupation of the built environment still demands analysis as well as political, social, and cultural consideration, particularly in China. To reflect upon these developing transformations, we have documented the emergent “automated landscapes”, spatial configurations and conditions engendered and afforded by automation, in three factories in the Pearl River Delta Region. These factories offer architectures where automation has been implemented at different scales and with different scopes of human and robot involvement in the workplace, including the integration of AI into the production line.
AUTOMATION FOR INCREASED RESPONSIVENESS
In 2005, Rapoo Technology, a Shenzhen-based electronics manufacturer of computer mice and keyboards, started to consider automation as a solution to labour shortages during peak seasons. It took the company four years to automate one of the production lines. By then, the product for which this technology was implemented was already outdated. As a result, Rapoo changed its approach and introduced a new, so-called soft manufacturing system based on a collaborative production model between humans and robots. Robots perform the relatively simple, repetitive tasks that include a higher volume of production of standardized components with a longer product run, as well as the dangerous and heavy tasks, while human workers carry out the functions that require more flexibility and that are, therefore, difficult to automate. This model, the company claims, brings resilience to the production line, allowing it to respond to future changes in product design and mass production.
Rapoo’s newly built factory in Pingshan District clearly manifests this collaborative system: different spatial configurations of human-machine relationships are organized in rectangular work islands, where levels of automation range from manual to fully automated manufacturing. Yet even in the most automated processes, human employees are still involved as supervisors, ensuring quality control in a process fully run by robots (see image 2, left). In most cases, manufacturing is semi-automated such that humans are assisted by robotic workers to perform their tasks, and vice versa (see image 2, middle). Interestingly, work islands for manual manufacturing follow the same spatial logics as the fully-automated counterparts, which would allow the factory to automate these tasks at a later stage (see image 2, right). Currently, there are around one hundred robots operating on various production lines in the factory and the number of manual labourers has declined from around 3,000 to approximately 700.
Aside from producing computer peripheral technologies, Rapoo has expanded its business to consult other companies on the implementation of automation technologies in their processes and production lines. Yet, for Rapoo, a fully-automated factory model is not a priority. According to its vice president, Steven Lee, the aim of automation is not to replace human labour but to increase the responsiveness of the factory. Human workers, Lee claims, are still more flexible than robots and will always be necessary in order to respond to market changes. (12)
The next step is to upgrade the factory towards smart manufacturing based on big data and AI, making the factory suitable to respond to consumer demands instantly. With the development of this direct link between factory and consumer, Rapoo expects to further reduce its workforce to 200 workers. This process of human replacement has also an influence on the design and program of the factory. As worker dormitories are no longer needed, one of the buildings has been reprogrammed a guest house for visiting engineers, or peers from other regions, while the other hosts workers in increasingly large rooms.
AUTOMATING THE HUMAN WORKER
Ash Cloud produces mobile phone and tablet accessories in an environment where the human and the robot work together. In order to attract factory workers in an increasingly competitive labour market, the company promotes the introduction of automation as an instigator of better working conditions—robotic arms operate as extensions of the human body. For example, automated carts that play sounds to alert humans as they are approaching, deliver components from one production line to another and relocate shipping boxes between designated locations in the storage space (see image 4, left). Yet, Ash Cloud produces more than fifty different products per day at varying volumes, which complicates the introduction of autonomous processes. The factory still depends on human labour for many assembly processes and maintains a spatial organization based on assembly lines operating during fixed hours (see image 3).
Automation, nevertheless, plays a pivotal role in increasing efficiency and productivity, in this case not by replacing human labour for assembly processes, but rather in the management of the human worker and the factory. In Ash Cloud, it is the Enterprise Resource Planning (ERP) system who manages the factory’s operation. The ERP, accessed in real-time by an IOS application on mobile devices, brings together size of supply and demand, product stocks, use of resources, working hours and holidays of employees, waste produced, as well as currency and stock exchange figures (see image 4, middle). Every process is monitored, every product and assembly part is coded, and every worker is individually managed, allowing the factory to be controlled internally and remotely. Screens, tablets, mobile phones, and other interfaces and forms of display measure and visualize performance. Worker productivity is tracked against targets and motivational illustrations—such as a sad turtle or a happy rabbit—command every assembly line (see image 4, right). Communication between the management team and the employees takes place via the app, making the middle manager redundant. The workers are trained for their specific tasks in manufacturing by watching the video embedded in the app displayed on the large screens in production line. The request for leave and other appeals is also mediated by the app, hence reducing the interaction with the higher management. In Ash Cloud, robots and AI have not replaced humans, they have become their managers. Human workers, in turn, are further automated.
TOWARD FULL AUTOMATION
Shenzhen based company Build Your Dreams (BYD) is the fifth largest battery producer in the world, and their electric vehicles are the foundation of Shenzhen’s public and private transportation systems. In 2014, the company opened an almost fully automated battery factory in Pingshan. While the outer shell of the building has not been altered in the transition to automation, the interior was redesigned to accommodate new environmental, safety, and security standards. BYD managers undergo six months of training in the control of production lines populated by robots. In addition to detailed manuals displayed on the walls, architectural elements also delineate the relations between different types of workers. For instance, glass walls marked with stripes just below human eye level, meant to prevent workers from hitting the glass, separate humans and machines both within the production zone and from the larger circulation areas.(see image 5). Engineers and even managers do not cross into the workers’ production zone. Instead, software and computer systems follow the entire process and, through QR codes, track the status and location of the newly assembled batteries along the production line. As humans disappear from the manufacturing process, productions lines can be stacked closer, and assembly and storage areas are reorganized to maximize productivity in the factory.
The system makes indoor work environment quieter less dusty, and void of human bodies. Its Research & Development division relentlessly works on further automating the factory. According to Jiang Shan, a higher BYD coordinating engineer, this level of automation has been possible and profitable only because battery production includes monotonous and predictable operations. (13) Worker housing has also been partially eliminated with this transformation and replaced by more space for manufacturing and, therefore, more production capacity. The company intends to open a new double-capacity, fully automated factory in Northwest China.
POSSIBLE AND DESIRABLE FUTURES
While in many other parts of the world automation seems synonymous with a future of work without workers, China has found alternative ways of integrating automation technologies into manufacturing and has thus become a site of experimentation for new modes of cooperation between humans and machines. With a yearly labour supply deficit of twenty-five percent, the Pearl River Delta region remains in constant need of workers.
Ash Cloud, BYD and many new factories in the region reconceptualize the idea of a production line as a living organism by combining human and robot productivity. AI takes over the management of tasks and turns for human workers, managing their efficiency and training them on how to do certain tasks, while simultaneously operating the work and maintenance of their robot collaborators. The real-time and online control of the human worker is matched to the level of detail of each component they produce. The changing relation between worker and machine seems to lead to a condition in which the worker is not anymore the main actor in the production process. As new types of workers are the result of the introduction of automation. Yet the larger impact of this seemingly safe and comfortable work environment, created by their robotic collaborators, is unknown. As those once in control of machines are repositioned and repurposed within the production cycle, they also have to renegotiate their place in the so-called smart city and society.
These automated landscapes and the assemblages of humans and machines they bring about are continuously redesigned—but not often by architects. Even if these technological systems demand and result in architectural and territorial transformations, from the scale of the factory to that of the city, and potentially the territories of Special Economic Zones, the domain of automation still lacks a critical spatial perspective. New configurations primarily reflect the requirements of the automation technologies, the efficiency required for productivity, corporate interests, and government incentives. However, the introduction of these technologies inevitably opens the doors to forms of spatial violence and segregation, expressed in reorganization of the industrial capitals and their peripheries into testing beds of The Next Economy, new economies integrating automation and supporting creative industries—such as in the case of Shenzhen. These transformations in the city’s labour market as well as the shift in the relations between labour, society, and architecture, materialize in developments to accommodate high-skilled workers and middle- and upper-income households, which is generally accompanied by the displacement of other sectors of society.
Shenzhen, having recently set its goal to become a creative city, has already been displaying the symptoms of a post-industrial city: mass production moving to the peripheries and other cities, rising real estate prices, new technologies and processes of gentrification. In Shenzhen the latter are strongly driven by urban redevelopment policies, aimed at attracting upper middle-class households. Its urban renewal agenda is mainly focused on the restructuring of so-called urban villages: dense informal settlements inhabited by millions of rural migrant workers, without access to social housing and security of tenure. (14) The redevelopment of these neighbourhoods have already led to the displacement of hundreds of thousands of low-income residents. (15) Many of them are forced to move, while others simply cannot afford the higher rents for their renovated homes and eventually relocate to other areas, farther away from their workplaces. Many Foxconn workers are facing the same problem. In an open letter, they have requested a salary raise to cover the higher rental costs in the district surrounding the factories. (16) It seems that even before their work is replaced by machines, they are being pushed out of their neighbourhoods to make room for a new group of highly-skilled workers.
At the same time, architecture, a discipline that has historically positioned the human body as its main reference, is now confronted with the possibility of shifting and widening its own focus to a post-human one. Critically analyzing these post-human environments and sites of encounter and collaboration between humans and non-humans is only a first step for exploring and enabling possible and desirable futures—futures where bodies, whether we refer to the human or robotic body, engage in non-exploitative forms of cohabitation.
The research on the Pearl River Delta region is conducted in close collaboration with Future+ Aformal Academy, and supported by Design Trust Hong Kong and the Consulate General of the Kingdom of the Netherlands in Guangzhou and Hong Kong. Research Team: Marina Otero Verzier (Director of Research), Marten Kuijpers, Klaas Kuitenbrouwer, Ameneh Solati, Ludo Groen, Víctor Muñoz Sanz, Grace Abou Jaoudeh, Emma Paola Flores Herrera, and Chris Zogopoulos. In collaboration with Merve Bedir, Jason Hilgefort, Junwen Wang, and Lucy Xia (Future+ Aformal Academy).
(1) Tania Branigan, “Taiwan iPhone manufacturer replaces Chinese workers with robots,” The Guardian, August 1, 2011, https://www.theguardian.com/world/2011/aug/01/foxconn-robots-replace-chinese-workers.
(2) International Federation of Robotics (IFR), Executive Summary World Robotics 2017 Industrial Robots, 2017, https://ifr.org/downloads/press/Executive_Summary_WR_2017_Industrial_Robots.pdf.
(3) Yu Huang and Naubahar Sharif, “From ‘Labour Dividend’ to ‘Robot Dividend’: Technological Change and Workers’ Power in South China,” Agrarian South: Journal of Political Economy 6, no. 1 (2017): 54, http://journals.sagepub.com/doi/pdf/10.1177/2277976017721284.
(4) Made in China 2025, http://english.cntv.cn/special/madeinchina/index.shtml
Scott Kennedy, "Made in China 2025," CSIS, June 1, 2015, accessed February 1, 2019, https://www.csis.org/analysis/made-china-2025
Find more information about ‘Made in China 2025’:
"Made in China 2025(中国制造 2025)", July 7, 2015, accessed February 1, 2019, http://www.cittadellascienza.it/cina/wp-content/uploads/2017/02/IoT-ONE-Made-in-China-2025.pdf (English translation of Made in China 2025)
(5) Huang and Naubahar, “From ‘Labour Dividend’ to ‘Robot Dividend’: Technological Change and Workers’ Power in South China,” 55.
(6) 南城经济科技信息网 “Management Measures for Special Funds for "Machine Substitution" in Dongguan”, March 25, 2016, accessed November 4, 2018,
“Dongguan City Promotion Enterprise ‘Machine Substitution’ Action Plan (2014-2016),”
Dongguan City Government Office, August 11, 2014, accessed November 4, 2018, http://www.dg.gov.cn/007330010/0202/201610/b07a766c890749519e9aae37cc741efb.shtmHuang Yu. Can Robots Save Dongguan? HKUST Seminar, March 16, 2017
(7) He Huifeng, “Foxconn’s Foxbot army close to hitting the Chinese market, on track to meet 30 per cent automation target,” South China Morning Post, July 1, 2015, http://www.scmp.com/tech/innovation/article/1829834/foxconns-foxbot-army-close-hitting-chinese-market-track-meet-30-cent.
(9) Dieter Ernst, “Advanced Manufacturing and China’s Future for Jobs,” East-West Center Working Papers: Innovation and Economic Growth Series, no. 8 (2018): 20.
(10) Huang and Naubahar, “From ‘Labour Dividend’ to ‘Robot Dividend’: Technological Change and Workers’ Power in South China,” 55.
(11) Boy Lüthje, “How Will China’s Industrial Modernization Plan Affect Workers?,” Huffington Post, October 17, 2017, https://www.huffingtonpost.com/entry/how-will-chinas-industrial-modernization-plan-affect_us_59e6aa04e4b0153c4c3ec3c4?guccounter=1.
(12) Steven Lee, interview by Merve Bedir and Marten Kuijpers, Shenzhen, China, 25 October 2017.
(13) Jiang Shan, interview by Merve Bedir, Pingshan, China, 25 January 2018.
(14) Ying Liu, Stan Geertman, Frank van Oort and Yanliu Lin, “Making the ‘Invisible’ visible: Redevelopment-induced Displacement of Migrants in Shenzhen, China,” International Journal of Urban and Regional Research, volume 42, issue 3 (2018): 488, https://onlinelibrary.wiley.com/doi/abs/10.1111/1468-2427.12646
(15) Ying Liu, Stan Geertman, Frank van Oort and Yanliu Lin, “Making the ‘Invisible’ visible: Redevelopment-induced Displacement of Migrants in Shenzhen, China,” International Journal of Urban and Regional Research, volume 42, issue 3 (2018): 490, https://onlinelibrary.wiley.com/doi/abs/10.1111/1468-2427.12646
(16) China Labour Bulletin, “Foxconn workers are the latest victims of Shenzhen’s gentrification”, China Labour Bulletin, 31 July 2018, https://clb.org.hk/content/foxconn-workers-are-latest-victims-shenzhen%E2%80%99s-gentrification (last access 24 December 2018)
Merve Bedir graduated from METU (B.Arch), and TUDelft (Ph.D). She is currently adjunct assistant professor in Hong Kong University, Department of Architecture. Merve is the co-founder of Land and Civilization Compositions (Hong Kong) and Aformal Academy (Shenzhen); and founding member of Kitchen (Gaziantep) and Center for Spatial Justice (Istanbul). Her work was reviewed in Metropolis, Avery Review, and The Guardian. She published in Thresholds, AD Magazine, and Volume, among others.
Marten Kuijpers is an architect based in Rotterdam and senior researcher at Het Nieuwe Instituut. His current research focuses on the spatial implications of automation, as part of the research project Automated Landscapes. Marten was curator of several exhibitions at the institute, including Sicco Mansholt. A Good European (2014) and Munich 1972 (2016). Between 2010 and 2013, he curated the lectures and debates program of the Netherlands Architecture Institute.
Marina Otero Verzier is a PhD architect, Director of Research at Het Nieuwe Instituut, and curator of “WORK, BODY, LEISURE,” the Dutch Pavilion at the Venice Architecture Biennale in 2018. With the After Belonging Agency, Marina was Chief Curator of the Oslo Architecture Triennale 2016. Previously, she was based in New York, where she was Director of Global Network Programming at Studio-X, Columbia University. She teaches at RCA in London.