When Robots Let Workers Rest
AI and Employment in China
It’s a sweltering summer day on a Shanghai construction site, the temperature soaring above 40°C. High above the ground, workers navigate steel frameworks, bending and tying rebar into place—a task that is exhausting, dangerous, and relentless. The sharp edges of the metal pose a constant threat; one misstep could mean a fall, a gash, or heatstroke under the merciless sun. For decades, this has been the reality of construction labor in China: physically demanding work performed in conditions that push the human body to its limits.
But the scene is changing. Shanghai Construction Group has deployed over 40 types of construction robots across its sites. Automated rebar-tying machines now save 50% of the labor previously required while boosting workflow efficiency by 10%. The workers who once toiled under that brutal sun? Many have been retrained as robot operators or maintenance technicians—safer jobs, with better prospects and higher skill requirements.
This story raises a fundamental question that’s often overlooked in Western discourse about “AI taking jobs”: What kinds of work should be replaced by machines? Are we asking the right question when we worry about “job displacement”—or should we be asking which jobs are worth preserving in the first place?
Jobs Humans Shouldn’t Do: The Human-Centered Case for AI
The most significant impact of AI on employment in China is not about “stealing jobs.” It’s about liberating workers from dangerous, grueling, and repetitive labor—the kind of work that, quite simply, humans shouldn’t have to do. This represents a fundamentally different technological philosophy: one where technology serves human emancipation rather than merely capital accumulation.
Shanghai Construction Group’s robotics fleet illustrates this principle in concrete terms. The company has developed over 40 types of specialized construction robots, each designed to take over tasks that have historically been hazardous or physically punishing. The underground diaphragm wall rebar cage welding robot, for instance, automates the precise positioning and welding of reinforcement bars, improving efficiency by 30%. The intelligent tunnel segment leveling robot reduces the need for manual labor by 50% while increasing production line efficiency by 10%. The automated welding robot cuts welding time by 25%—time previously spent by workers in cramped, poorly ventilated spaces, exposed to fumes and intense heat.
These robots share a common characteristic: they replace work that is dangerous, exhausting, and repetitive. They take over tasks performed in scorching heat, at dizzying heights, or in toxic environments—conditions where prolonged human exposure carries serious health risks. This isn’t about replacing people; it’s about redefining what work should demand of the human body.
The safety improvements are quantifiable and dramatic. Global research on robotics deployment has found that for every additional robot per thousand workers, workplace accidents decrease by 0.254 incidents and fatalities drop by 0.0353 deaths. In China specifically, the deployment of robots between 2013 and 2019 led to a 68.5% reduction in workplace accidents and a 71% decrease in fatality rates. These robots are now performing critical functions in five major high-risk sectors: mining, bomb disposal, chemical handling, construction, and manufacturing.
The contrast with Western approaches to AI and employment reveals a deeper philosophical divide. In much of the Western discourse, the conversation centers on metrics like unemployment rates, income losses, and economic efficiency. The underlying assumption is often that technological advancement is primarily about optimizing capital returns, with labor protections and social safety nets added as afterthoughts—mechanisms to manage the “disruption” rather than to guide it.
China’s approach reflects a different set of priorities. The focus is on workplace safety, improvement of working conditions, and the principle of putting people first. Policy responses emphasize proactive training, rights protection, and career development pathways. The typical case study differs accordingly: where Western narratives might feature warehouse workers anxious about being replaced by automation, Chinese narratives highlight construction workers being freed from hazardous positions and gaining access to skill upgrades.
This philosophical difference is codified in national policy. The reform plan for building a high-quality industrial workforce, issued by the central government, explicitly emphasizes “strengthening workplace safety and occupational disease prevention.” The vocational skills training program, targeting 30 million participants between 2025 and 2027, allocates 60% of corporate education budgets to frontline workers. The “New Eight-Grade Worker” system establishes clear pathways for skill advancement, with targets to cultivate approximately 2,000 national master craftspeople, 10,000 provincial master craftspeople, and 50,000 municipal master craftspeople by 2035.
The underlying logic is clear: this isn’t simple labor replacement. It’s workforce transformation—a process of upgrading human capabilities while machines take over the tasks that should never have been human work in the first place.
The Demographic Dimension: Why China’s AI Story is Different
China’s working-age population is undergoing a significant, long-term contraction. According to projections from the National Information Center, the labor force will decline from 879 million in 2020 to 787 million by 2035—a reduction of 92 million people, or an average annual decrease of 5.91 million workers. By 2050, this figure is expected to fall further to 601 million.
The drivers of this trend are structural and irreversible. In 2023, China recorded just 9.02 million births—the seventh consecutive year of decline. The total fertility rate has fallen to 1.02, far below the replacement level of 2.1. Surveys indicate that post-90s cohorts plan to have an average of 1.54 children, while post-00s cohorts plan for 1.48. These aren’t temporary fluctuations; they represent a fundamental demographic shift that will define China’s labor market for decades to come.
Against this backdrop, the scale of AI deployment takes on a different meaning. The construction robotics market delivered approximately 10,000 units in 2023, with BRIGHT DREAM ROBOTICS alone having deployed over 3,000 construction robots covering about 20 million square meters of construction area. The market is projected to reach 1.92 billion yuan by 2029, with an average annual growth rate of 15%. Meanwhile, labor productivity has risen from 123,248 yuan per person in 2018 to 161,615 yuan per person in 2023—a compound annual growth rate of 5.57%.
The crucial point is this: the scale of AI labor substitution aligns with the scale of labor force contraction. Rather than creating a jobs crisis, AI is filling a gap that would otherwise constrain economic development. This represents a form of “temporal alignment”—a rare convergence where technological capability matches demographic necessity.
The contrast with developed Western economies is stark. While the International Monetary Fund reports that AI will affect 40% of jobs globally, with 60% of jobs in advanced economies facing greater impact, the nature of that impact differs fundamentally depending on labor supply trends.
This distinction is critical for any country in the Global South attempting to understand how AI will affect its own labor market. Whether AI represents a threat or an opportunity depends heavily on domestic demographic structure and labor supply trends. For countries facing labor shortages, AI becomes a reasonable—even necessary—solution rather than a disruptive force. For countries with labor surpluses, a more cautious approach with different policy interventions may be required.
The Grassroots Enthusiasm: AI as Empowerment, Not Displacement
The story of AI and employment in China has another dimension, one that complicates any simple narrative of labor replacement. While urban professionals worry about whether AI will render their skills obsolete, something very different is happening in China’s less-developed regions: AI tools are creating economic opportunities where few existed before.
Consider Rongjiang County in Guizhou Province, a mountainous area in China’s southwest. In 2023, this small county became nationally famous not for technology, but for soccer—specifically, the Cun Chao (Village Super League), a grassroots sports tournament that captured national attention. What’s remarkable isn’t just the tournament itself, but the digital economy that emerged around it.
Leveraging the attention generated by the Village Super League, Rongjiang cultivated over 2,200 local livestreaming and online marketing teams. The county now has more than 12,000 new media accounts actively promoting local agricultural products and cultural experiences. These efforts generated online and offline sales of agricultural products totaling 146 million yuan. During the 2023 National Day holiday alone, the county received 498,900 tourists, producing tourism revenue of 602 million yuan—a year-on-year increase of 21.99%.
The significance of this case extends beyond the impressive numbers. The AI-powered tools that enable livestreaming, video editing, automated translation, and e-commerce logistics have made it possible for farmers and small entrepreneurs—people who never had access to professional broadcasting equipment or marketing expertise—to participate in the digital economy. These tools lower barriers to entry,
This pattern is repeating across China’s grassroots governance and service systems. In Beijing’s Haidian District, the “GuanXin” AI system has generated over 530,000 visits and nearly 80,000 question-and-answer interactions, helping residents navigate government services. In Taiyuan’s Xiaodian District, the “Shutong Xiaodian” platform has integrated DeepSeek-R1, a large language model, to improve public service delivery. In Hangzhou’s Shangcheng District, an “AI grid administrator” connects 16 community WeChat groups, with a knowledge base covering 65 hot-topic issues. Similar applications are being deployed in the Guangdong-Hong Kong-Macao Greater Bay Area’s intelligent health support network and Shanghai’s Putuo District micro-governance system.
The common thread in these deployments is high acceptance and enthusiasm at the grassroots level. Unlike urban elites who may view AI with apprehension, grassroots users often embrace these tools because they experience AI as empowerment rather than disruption.
This enthusiasm is mirrored in the explosive growth of flexible employment. China now has 200 million people engaged in flexible work arrangements, representing 27% of the total employed population. The proportion of flexible employment positions in recruitment has surged from 8.4% in 2019 to 15.2% in 2024. Platform-based economies have created vast numbers of jobs in ride-hailing, delivery services, livestreaming, and content creation—occupations like “AI trainer” and “data labeling specialist” that didn’t exist a decade ago. These roles have lower barriers to entry and offer greater flexibility than traditional employment.
The geographic and economic divide in responses to AI is revealing. For urban middle-class professionals with established careers, AI can feel like a threat—a technology that might devalue their expertise or automate their functions. For grassroots communities and individuals who previously lacked access to resources and opportunities, AI represents something quite different: a chance to participate in economic activities that were previously out of reach.
The Systemic Response: From Policy to Philosophy
In September 2024, the Central Committee of the Communist Party of China and the State Council issued the first comprehensive strategic document on employment in the new era, outlining 24 measures covering macroeconomic regulation, industrial coordination, skills training, support for key demographics, public services, rights protection, and organizational leadership. Significantly, Article 23 explicitly addresses “actively responding to the employment impacts of rapid development in emerging technologies such as artificial intelligence.”
The policy framework introduces several institutional innovations. It establishes a unified national employment information resource database to improve labor market matching. Crucially, it incorporates employment outcomes into the performance evaluation system for county-level and higher Party and government leadership—a mechanism that ensures local officials prioritize job creation and workforce transition alongside economic growth metrics.
This isn’t an after-the-fact remediation strategy; it’s proactive guidance combined with comprehensive support.
The skills training component is particularly ambitious in scale. Between 2025 and 2027, China plans to provide subsidized vocational training to over 30 million people, with a focus on digital economy sectors including AI, big data, advanced manufacturing, low-altitude economy, transportation, agriculture, and lifestyle services. The training system employs the “New Eight-Grade Worker” framework, which creates clear skill progression pathways. Enterprises are required to allocate 60% of their education budgets to frontline workers, ensuring that those most affected by technological change receive priority in skill development.
The goal isn’t merely employment; it’s what policymakers call “skill-based employment, skill-based income growth, and skill-based fulfillment”—a formulation that reframes work as a source of capability development and personal satisfaction, not just income.
Regional balance is another pillar of the systemic response. Ten ministries, led by the Ministry of Industry and Information Technology, have issued guidance on promoting the transfer of labor-intensive industries from eastern coastal regions to the central and western interior. The goal is for these regions to significantly enhance their industrial absorbing capacity by 2025. The logic is straightforward: as AI frees up labor in the east, facilitating the return of workers to their home regions—combined with industrial relocation—can support county-level economic development and avoid the regional imbalances that plagued earlier waves of industrialization.
Mechanisms include sharing arrangements for output, revenue, and land use indicators between sending and receiving regions, creating aligned incentives for cooperation.
Perhaps most intriguing is the long-term theoretical framing. Scholars working within a Marxist framework have articulated a vision in which AI’s productivity gains could eventually support a transformation in the nature of labor itself. In the advanced stage of communism, as originally theorized, labor would transition from being primarily a “means of livelihood” to becoming “the foremost necessity of life”—work as a form of self-realization rather than mere survival.
Beyond Replacement: A Framework for Understanding
China’s experience with AI and employment reveals something crucial: the impact of AI on work is not determined by the technology itself, but by three variables—demographic structure, value choices, and institutional design. Countries facing labor shortages will experience AI differently than those with labor surpluses. Societies that prioritize worker safety will deploy AI differently than those that prioritize cost minimization. Political systems with strong capacity for coordinated intervention will manage transitions differently than those relying primarily on market mechanisms.
For countries across the Global South, the lesson isn’t to replicate China’s specific policies. It’s to understand how these variables operate within your own context. What is your demographic trajectory? What values will guide which jobs you automate and which you preserve? What institutional mechanisms can you deploy to ensure that technological gains are broadly distributed rather than narrowly captured?
The Rongjiang case offers a particularly relevant insight: you don’t need Silicon Valley infrastructure or massive capital to participate in the AI era. Creativity, local knowledge, and access to increasingly democratized digital tools can be enough—especially for communities that are starting from a position where even basic opportunities have been scarce.
Those workers on the Shanghai construction site—the ones freed from tying rebar in 40-degree heat—are now training to become robot technicians, learning to operate and maintain the machines that have taken over the dangerous work. This isn’t an ending of their career, but a new beginning.
Thanks for writing this, it clarifies a lot, truly making me wonder, like after a focused pilates class, about AI's potential to liberate us from work humans shouldn't do.