Human Cost of Faster Work: Chinese Employees Weigh AI’s Promise Against Pay Fears

Human–AI collaboration remained the dominant approach across the enterprises studied, with businesses building workflows that still depended on human judgement and oversight.

Artificial intelligence is helping Chinese businesses handle more customer queries, recruit staff faster and improve factory output, bringing changes that employees can already see in their daily work. New research from the International Labour Organization (ILO) finds that these gains come with concerns about job security, uneven access to skills and future earnings. For workers, the question extends beyond whether AI can make their jobs easier to whether they will have the training and support to benefit from the changes it brings.

The ILO research brief, Artificial Intelligence Adoption in Chinese Enterprises: Productivity Effects, Workforce Implications, and Policy Challenges, was prepared by Ekkehard Ernst of the ILO Research and Statistics Department with researchers at Renmin University of China. It draws on interviews with 21 enterprises and a survey of 1,591 professionals, covering manufacturing, finance, business services, construction, education, media and travel. Businesses ranged from an eight-person start-up to a conglomerate employing 270,000 people, with every interviewed firm already using AI or making concrete plans to adopt it.

Faster Work Produces Striking Results

At one insurance company, 300 customer-service employees increased the number of issues handled each day from 6,000 to 15,000. A large insurance group cut its recruitment cycle from 30 days to 13 days, a reduction of around 57 per cent, and a smart manufacturing facility reported a 30 per cent improvement in production efficiency. These examples show how AI can change the amount of work teams complete and the time required for familiar business processes.

Those figures require careful reading because they were reported by the firms themselves and have not been independently verified. The 21 enterprises were deliberately selected for the study, meaning their experiences cannot be statistically generalized to all Chinese businesses. Most participating firms also lacked systematic ways to assess AI's effects, particularly changes beyond conventional productivity measures. Counting completed tasks or days saved leaves unanswered questions about job quality, working conditions and how employees' responsibilities are being reshaped.

Workers See Opportunity and Reasons for Concern

Among the professionals surveyed, 56 per cent viewed AI adoption as inevitable and 47 per cent believed it would create more jobs than it displaced, reflecting greater optimism than surveys in many OECD countries have found. Expectations about personal earnings were less reassuring, with 39 per cent anticipating income declines. The reported productivity gains were concentrated in repetitive, data-intensive activities such as document processing, customer-query handling, résumé screening and data collection, placing routine clerical, administrative and customer-service roles under particular pressure.

Human–AI collaboration remained the dominant approach across the enterprises studied, with businesses building workflows that still depended on human judgement and oversight. Putting those arrangements into practice brought difficulties involving employee resistance, skills shortages, unreliable AI output, data security, regulation and integration with existing systems. An education-technology company identified a marked gap in AI capabilities among employees over 40, and other firms described broader weaknesses in AI literacy and employees' ability to use the tools effectively.

Training and Support Will Shape Who Benefits

Ernst described the central change as businesses reorganising work around people and AI, with skills and management capacity representing a bigger bottleneck than the technology itself. The brief identifies four priorities: strengthening AI skills and lifelong learning, particularly for mid-career and older workers; helping employees move into higher-value tasks; improving measurement of productivity, job quality and working conditions; and giving smaller firms access to shared platforms, training and affordable AI services. These measures would help turn business efficiency gains into decent work and spread the benefits more broadly across workers and enterprises.