Historically, the United States and China grew rich by producing a wide range of goods for global markets. To stay rich, they are both turning to artificial intelligence (AI) and associated technologies to sustain their manufacturing market share. The two countries recognize that AI-enabled manufacturing could accelerate production, reduce downtime and waste, improve production capacity, reduce labour costs and optimize supply chains. However, policy makers in China and the United States aren’t prepared for how these technologies may alter the global economy and potentially lead to manufacturing overcapacity — too many goods facing too little demand.
In 2015, Chinese officials determined that the country could not depend on its traditional strategy of cheap labour coupled with generous subsidies to remain the world’s foundry. Instead, they bet on automation. China was willing to accept higher levels of unemployment and labour unrest to maintain its competitiveness by relying on AI expertise and a robot army. Today, China has the world’s largest manufacturing workforce and is the largest deployer of industrial robots — 295,000 units operating with limited human supervision.
But the United States and China are not alone in using AI and associated technologies to bolster manufacturing. Japan, Germany and South Korea (among others) have also developed programs to encourage AI-enabled manufacturing. At the national level, these efforts make sense, but, collectively, they could create an economic catch-22. In essence, AI-enabled manufacturing will help individual firms to directly respond to changing consumer preferences and market demand. But if every country has such firms, and these firms have invested in expensive AI manufacturing, they may need to sustain production and ignore market demand, which could lead to dumping and overcapacity.
AI-enabled manufacturing competition could have other important spillovers. First, AI-enabled manufacturing will alter which countries trade and how they produce goods. The World Trade Organization (WTO) warns that “AI could significantly reshape economies’ comparative advantages by shifting the relative productivity of labour, capital and knowledge across sectors and economies.” This, according to the WTO, would result in economies that may begin to specialize differently as AI transforms key industries, given that AI-driven productivity gains are likely to vary across sectors. The WTO also found that China’s growing use of AI-enhanced manufacturing may “redefine the comparative advantage of economies by reducing ‘reliance on labour’ to a ‘greater reliance on intelligence systems that can enhance productivity, efficiency and decision-making.”
Second, AI-enabled manufacturing will alter which firms dominate markets by providing new services associated with goods. But AI-enabled technologies are expensive, and only large companies can afford such investments. Companies that can link their manufactured goods to services associated with those goods can make more money from every sale.
For example, Boeing uses various AI systems, such as digital twins, not only to build planes but also to monitor the health of the planes it has sold. It offers its customers associated services such as basic maintenance, training and refurbishing. More recently, Boeing used AI-enabled manufacturing to deliver military aircraft to Australia. These planes use AI to reduce pilot risk and sensory overload and employ new approaches to combat. As such, Boeing has become more responsive to its customers, allowing it to further dominate the civilian and military markets.
Third, China may be the best-positioned country to sustain production in such conditions. Chinese officials have provided patient capital and a supportive policy environment for firms to utilize AI in their factories. Their research labs sit close to their factory floors — allowing companies to quickly test and improve AI-enabled manufacturing systems — and Chinese firms compete domestically to cut costs and eagerly embrace replacing labour with technology. China has a large supply of data, computing power, AI expertise and capital — key elements of the AI-enabled manufacturing supply chain. Several analysts have concluded that China is already the world’s leading integrator of AI.
Fourth, AI-enabled manufacturing could fuel social and political instability. Public investment in automation may be seen as favouring machines over workers, eroding public support for technological adoption and strengthening protectionist sentiment. Governments that fail to respond to these concerns may find themselves out of power.
Finally, when multiple countries simultaneously pursue domestic AI-enabled manufacturing capabilities, they may collectively create overcapacity in different manufacturing sectors. Policy makers are left to decide if supporting unsuccessful firms to maintain capacity for national security purposes, prevent job losses, foster competition or prevent a recession is the right move.
AI is set to transform how, when and where we make things. Factories will become smarter, more flexible and more competitive. Nations that master AI-enabled manufacturing may dominate exports, while others face declining market share and rising pressure for subsidies and protectionism. For more than 50 years, policy makers have been unable to find shared solutions to reduce capacity or to coordinate output across key sectors, but given these stakes, policy makers in Beijing, Washington and elsewhere must collaborate now to mitigate potential economic and social disruption.
