Will artificial intelligence, hereinafter referred as AI, replace human labor? Whose jobs are most at risk, and how soon such changes might unfold?

Living in a revolutionary era along with AI, concerns about job displacement have become increasingly widespread. Yet the sub-forum: Labor Market Transformation in the Age of Artificial Intelligence: New Challenges for China and the World of Shanghai Forum 2026, chaired by Professor Zhang Jun, Dean of the School of Economics at Fudan University, brought together distinguished scholars from multiple disciplines to offer more nuanced perspectives.

The impact of AI is not best understood as simple job displacement in both China and other parts of the world, but as a structural transformation of how labor would be defined, organized, and valued in the future job market, both for the employees and employers. Besides, across different professions, AI affects different types of job in different ways: some tasks might be taken more easily than others, while the pace of job displacement also differs.

A Basic Question: Replacement or Reorganization?

The idea that AI will “replace jobs” assumes that work disappears in discrete units. But several speakers argued that this framing may be overly simplistic. AI, indeed, is reorganizing production, shifting tasks and redistributing income.

Drawing on his observation on different fields including STEM, clinical and social science research, Richard B. Freeman, Herbert Ascherman Chair in Economics at Harvard University, noted that large reshuffling of job tasks might soon happen as people adjust to the evolving world of work. He pointed out that humans should focus on areas where they still hold a comparative advantage, “Throughout history, technological fears of ‘the end of work’ have often turned into excitement about new possibilities.”

But, at the same time, Freeman admitted that new risks might also be raised from the trend. The use of AI-generated avatars of CEOs and scientists, usually appeared as cloned versions of real people, can lead to structural and legal problems which can only be resolved with joint efforts of policy makers.

The real question, therefore, is not whether jobs will vanish, but how the internal structure of work and job market are being reshaped.

AI Is Evolving: Data-driven generative models and AI agents Matters More

Lin William Cong, Presidential Chair Professor and Associate Dean at Nanyang Business School, Nanyang Technological University, approached the transformation from a technological and modeling perspective. He highlighted the shift from instruction-driven systems to data-driven generative models and AI agents, which are increasingly capable of simulating complex economic environments.

Cong introduced the idea of a “world model” in economics, where AI systems—integrating reinforcement learning and multi-agent simulation—can be used not only to assist decision-making, but also to model and potentially reshape firm behavior and market dynamics. At the same time, he noted that our current understanding of how these systems operate remains limited, suggesting that future research will need to further integrate economics with computational sciences to better understand human–AI collaboration in evolving labor systems.

What Is Actually Changing: From Tasks to Systems

At the core of this transformation is a shift in the whole labor system.

For employers, the trend of applying AI in work field means the urgency of building “AI-enabled organizational capabilities”—where value lies not just in simply applying AI, but in how to transform AI from digital tools to digital infrastructure.

ZHU Feida, associate professor of Computer Science, Singapore Management University, emphasized that the future intelligence systems will not rely solely on AI, but by the integration of human intelligence, artificial intelligence, and organizational intelligence.

While AI excels at scale, speed, and pattern recognition, human intelligence remains essential for perspective, creativity, and ethical judgment. The most effective systems, therefore, are those where AI and human are effectively distributed with different tasks, and workflows are structured so that the whole becomes greater than the sum of its parts of its parts.

ZHANG Dandan, professor in economics and deputy dean at the National School of Development, Peking University, introduced three methods of measuring the impact of AI on employment. The results revealed that there were already adjustments in entry-level hiring contracts due to the growth of AI models, and long-run structural risks of white-collar cognitive jobs are comparatively high compared to other jobs. Still, according to the latest measurement, there remains a gap between theoretical exposure and actual usage of AI replacement.

Why It Feels Like Disruption: Perception and Uncertainty

If the transformation is gradual, why does it feel so disruptive?

Part of the answer lies in perception. Research presented by Joonseook Yang, assistant professor from the Department of Political Science & International Studies, Yonsei University, shows that workers’ fears are not primarily driven by objective risk, but by information environments and expectations. In highly digitalized contexts, many already assume that disruption is inevitable.

This highlights a tension between perception and reality: even when actual job loss is limited, perceived insecurity can be widespread. Negative information reinforces existing anxiety only marginally, while reassurance can significantly shift attitudes. AI, in this sense, operates as both a technological and psychological force.

From a macroeconomic perspective, XIE Danxia, Associate Professor at the School of Social Sciences, Tsinghua University, examined how AI may reshape long-term growth dynamics and labor demand. She proposed a general analytical framework for what she termed a “digital-intelligent economy,” where production and innovation are increasingly driven by data, computing power, algorithms, and storage.

In extreme scenarios, such a system could significantly reduce reliance on traditional labor inputs. However, XIE emphasized that AI’s impact on employment is multifaceted: while some tasks may be replaced, new opportunities may emerge through enhanced innovation efficiency, reduced knowledge costs, and accelerated technological diffusion. She also pointed out that institutional responses—such as changes in working hours or labor regulations—may become central in shaping how societies adapt to these transformations.

LIAO Fangli, distinguished research fellow, China Development Institute (Shenzhen), highlighted three emerging challenges that may disrupt us in the AI era. These challenges are not merely constraints, but also opportunities that can be actively leveraged.

First, there is a growing risk of structural mismatch between education and labor market demand; second, the importance of interdisciplinary capabilities is rising, where technical knowledge must be combined with language skills, legal awareness, and ethical judgment, especially in globally competitive AI products; third, Liao points to the emergence of “super individuals”, whose productivity is significantly amplified by AI.

At the micro level, Hu Bo and Li Rui from the School of Economics at Fudan University presented empirical evidence on platform-based competition and its implications for labor and business sustainability. Using data from major food delivery platforms, they analyzed the effects of large-scale subsidy campaigns on restaurant operators.

Their findings suggest that while subsidies significantly increase order volume and platform traffic, they also tend to lower average transaction values and crowd out offline dining. As a result, overall revenue gains for merchants remain limited, and in many cases, net profits decline due to the erosion of higher-margin dine-in business. Smaller independent restaurants and those more heavily involved in subsidy programs appear particularly vulnerable, often facing a dilemma between losing orders and sacrificing profitability.

This case illustrates how digital platforms and algorithm-driven competition can reshape not only labor demand but also the broader economic environment in which businesses operate.The attendees also discussed what broader questions such shift can raise: how income is distributed, how productivity gains are shared, and how social systems adapt when work is no longer the primary anchor of economic life.

The discussions at the sub-forum suggest that the impact of AI extends beyond job displacement, pointing instead to a broader transformation in how work is organized and valued. AI is not simply removing jobs—it is redefining what work is, how it is organized, and where its value lies. The future of labor will likely not be jobless, but it will be less certain, less familiar in required skills, more about building collaboration with AI instead of simply using them.

In that sense, the real challenge is not technological substitution, but also institutional and structural adaptation—how societies respond to a world where work is no longer what it used to be.

The sub-forum was part of the Shanghai Forum 2026, a three-day annual conference under the theme “The Age of Reconfiguration: Innovation and Co-governance”. Co-hosted by Fudan University and the Chey Institute for Advanced Studies, this year’s Shanghai Forum brought together nearly 400 participants from over 50 countries and regions for discussions spanning artificial intelligence governance, green transition, and the Global South development.

(END)

Writer: YANG Xinrui

Editor: WANG Mengqi, LI Yijie