Advancing social justice, promoting decent work
ILO is a specialized agency of the United Nations
English
Research brief
Artificial intelligence adoption in Chinese enterprises: Productivity effects, workforce implications, and policy challenges
This research brief explores the adoption of artificial intelligence in Chinese enterprises and its implications for productivity, employment and the world of work.
AI adoption is pervasive across Chinese enterprises of all sizes. Across the 21 firms studied through in-depth interviews spanning manufacturing, finance, business services, construction, education, media, and travel, every company reported either active AI deployment or concrete plans for imminent adoption. A complementary survey of 1,591 professionals found that 56 per cent view AI as an inevitable trend, while 47 per cent believe AI creates more jobs than it displaces.
Productivity gains are substantial where measured, but measurement itself remains patchy. Reported improvements include a 57 per cent reduction in recruitment cycle time, a 150 per cent increase in daily customer-query throughput, efficiency gains of 30–100 per cent across core workflows, production efficiency improvements of 20–30 per cent in smart manufacturing, and labour-equivalent savings of 5–6 full-time employees from AI data handling. Yet most firms lack systematic evaluation frameworks.
Enterprises adopt AI through three distinct organisational models: centralised specialist teams (common in technology firms), business-embedded integration (prevalent in customer-facing industries), and organic bottom-up diffusion (typical of smaller firms). Productivity impact deepens as firms progress from tool-assisted efficiency through process-embedded optimisation to business-model innovation.
Workforce implications are significant: employee resistance, age-related digital divides, and displacement anxiety are widespread, with 39 per cent of surveyed professionals anticipating income declines. Firms report that AI is most effective in automating repetitive, data-intensive tasks, creating hybrid human–AI workflows rather than eliminating human oversight. Output quality limitations, regulatory constraints, and skills gaps remain the dominant barriers to deeper integration.
