Reshaping employment in the AI Era: Reinstatement or displacement?

Editor's note: Xu Haoqing is an associate research fellow at the Artificial Intelligence Economics Research Office, Institute of Economics, Chinese Academy of Social Sciences. The article reflects the author's views and not necessarily those of CGTN.

Artificial intelligence (AI) is profoundly reshaping the landscape of the labor market. From the launch of Tesla's autonomous taxi to the multi-city pilot operation of Baidu's Apollo Go, alongside the successive advent of large language models such as ChatGPT and DeepSeek, the impact of technology on traditional jobs has repeatedly become the focus of public debate.

Fears that technological progress will eliminate human jobs are far from new. Back in the first Industrial Revolution, the adoption of power looms and knitting machines triggered the Luddite Movement. Terrified by unemployment and skill obsolescence, workers protested by destroying machinery. In 1930, John Maynard Keynes coined the concept of "technological unemployment," predicting that this new ailment born of technological advancement would continue to emerge. 

Today, amid the explosive development of AI as a general-purpose technology, such concerns have not abated but grown more intense.

Nevertheless, a frequently overlooked truth remains: Displacement from existing jobs by technology does not condemn workers to permanent exclusion from the labor force.

Historically, automation is by no means a recent phenomenon; it stretches back at least 200 years. Yet, why have labor shares remained largely stable over these two centuries, with productivity growth keeping pace with wage growth? 

Economic research offers an explanation: the reinstatement effect. While technological progress phases out old tasks, it continuously generates new ones that create opportunities for workers to re-engage in production.

Agricultural mechanization serves as a typical example. Though it reduced labor demand for manual farm work, technological advances and emerging service needs gave rise to new forms of employment. 

This drove the labor market toward a new higher-level equilibrium and facilitated industrial-structure upgrading. Acemoglu and Restrepo's widely cited 2019 framework ("Automation and New Tasks: How Technology Displaces and Reinstates Labor") notes that technological change can raise labor demand by creating new tasks and jobs directly associated with new technologies. 

By boosting productivity, unlocking production resources and lowering production and consumer prices, automation can also stimulate employment indirectly through rising demand for goods and services.

A growing body of empirical evidence suggests that the reinstatement effect brought by automation is gradually outpacing the conventional displacement effect. Firms adopting new technologies often generate fresh employment as they scale up operations, even as non-automated competitors may be pushed out of the market.

We must soberly recognize, however, that this transition will not occur automatically. 

In the short run, technological change inevitably gives rise to frictional unemployment and skill mismatches. Firstly, workers displaced by automation need time to adapt to shifting skill requirements, and the mismatch between existing skills and technological demands prevents them from taking up new jobs immediately. 

Secondly, structural shifts in skill requirements mean many workers' existing competencies cannot be directly applied to emerging occupations. 

Empirical studies find that occupations with high AI exposure experience a notable drop in labor demand. New technologies may also widen within-occupation wage gaps and suppress wage growth. If adjustments are left entirely to market forces alone, job-wage polarization will intensify, and income and wealth inequality will widen. 

Theoretical analysis further indicates that although the net employment effect of AI hinges on the balance between substitution and reinstatement, without effective public policies, gains brought by AI accrue mainly to a small group of capital owners and technology controllers, which will exacerbate income concentration and inequality across the globe.

This topic is of special significance to China. As the country with the world's largest labor force, China is in a critical period of advancing high-quality development through industrial upgrading and economic transformation. Research on the potential impacts of AI on the labor market is therefore highly urgent in practical terms.

Meanwhile, relatively high educational attainment and extensive digital technology application endow China's workforce with unique comparative advantages for adapting to the new economy. 

How to translate such technological and educational strengths into workers' capacity to adapt to the new economy – enabling workers not merely to survive in the AI era, but to excel in new tasks and positions – constitutes a vital agenda for unlocking AI's employment potential and achieving high-quality employment.

Public policies must play an active role in this transformation: The reinstatement effect of technological change should be effectively translated into sustainable job opportunities to facilitate workers' transition from old posts to new ones. 

Firstly, scale up vocational training and re-employment programs. Job transitions may bring short-term unemployment pressure and skill-matching problems. Hence, targeted retraining should be provided for displaced workers to help them quickly meet the requirements of new jobs and mitigate structural unemployment. 

Secondly, improve skill-certification systems and employment services. By promoting vertical skill upgrading and sufficient cross-industry labor mobility, we can efficiently align workers' competencies with dynamic market demands and offer institutional support for workers during job transitions. 

Thirdly, foster and develop emerging industries. Disruptive innovation driven by artificial intelligence can open new realms of economic activity and provide broad scope for labor reinstatement. 

Developing AI-complementary technologies that integrate human unique strengths with intelligent systems is crucial for addressing employment challenges.