Technological unemployment collides with artificial intelligence

Introduction: The Employment Effects of Technological Change

In October 2025, Amazon announced the layoff of 14,000 corporate positions, a decision that marks the beginning of a substantial impact of artificial intelligence technology on white-collar employment. The company's statement indicated that this organizational restructuring aims to optimize operational efficiency and reallocate resources to strategic areas such as generative artificial intelligence. This case reveals the intrinsic connection between technological advancement and the structural adjustment of the labor market, sparking a new round of discussions about technological unemployment.

The concept of technological unemployment was first proposed by Keynes in 1930, defined as the reduction in labor demand caused by technological innovation. Historical data shows that this phenomenon has obvious cyclical characteristics. According to bibliometric analysis, the term “technological unemployment” has three significant peaks in frequency of occurrence during the 1920s-1930s, the 1960s, and after 2010, corresponding to the technological diffusion periods of the Second Industrial Revolution, the wave of automation, and the artificial intelligence revolution.

Currently, despite the unemployment rate in the United States remaining at a relatively stable level of 4.3%, the structural adjustments in white-collar positions have drawn widespread attention. This article will explore the impact mechanism of artificial intelligence on the job market through historical comparative analysis, assess potential risks, and provide corresponding policy recommendations.

Historical Comparative Perspective

The experience of the Industrial Revolution indicates that the impact of technological advancement on employment has distinct structural characteristics. In the early 20th century, the annual growth rate of productivity in U.S. manufacturing exceeded 5%, but this growth was accompanied by a 20% decline in agricultural employment. During the period from 1929 to 1933, the unemployment rate rose from 3% to 25%, demonstrating that technological change may exacerbate employment pressures during economic downturns.

The wave of automation in the 1960s further confirmed this structural impact. Research at the time showed that the substitution effect of automation technology on manufacturing employment was significant, but due to the expansion of employment in the service sector and the special demand brought about by the Vietnam War, the overall job market remained relatively stable. During this period, the U.S. government established a special committee to study the impact of automation on employment, providing important references for subsequent policy formulation.

In the long term, the employment effect of technological progress depends on the dynamic balance between the substitution effect and the compensation effect. The substitution effect is reflected in the replacement of existing positions by technological advancement, while the compensation effect manifests in the creation of new positions and the demand growth brought about by a decrease in production costs. Historical experience shows that this balance requires appropriate policy intervention and the coordination of the market environment.

Economic Impact of Artificial Intelligence

On a macro level, artificial intelligence technology is becoming an important driving force for economic growth. Between 2023 and 2025, investments related to artificial intelligence are expected to contribute nearly 1 percentage point to the growth of the US GDP. Corporate profit margins have increased from 6.5% in 2003 to 10.69% in the second quarter of 2025, demonstrating the impact of artificial intelligence technology on improving production efficiency.

Industry-level data shows that the impact of artificial intelligence has significant heterogeneity. In the banking sector, AI technology has improved fraud detection accuracy to 95%; in the insurance industry, claim error rates have decreased by 20%; the energy sector has reduced equipment downtime by 30% through predictive maintenance; the retail sector has achieved a 15% increase in sales through personalized recommendations; and in healthcare, assisted diagnosis has improved medical efficiency by 25%.

The efficiency improvements are underpinned by profound adjustments in the employment structure. The layoffs at Amazon indicate that white-collar positions, such as management and data analysis, are facing direct impacts. The company plans to enhance the efficiency of middle management by 30%-50% through organizational flattening. This shift signals a fundamental transformation in the traditional model of knowledge work.

Characteristics of Employment Market Transformation

The transformation of the current job market exhibits the following notable characteristics:

First, the skill structure of affected positions has changed. Traditionally, automation technology primarily impacts programmed production jobs, while artificial intelligence technology can replace some non-programmed cognitive tasks. This means that traditional high-skill fields such as education, finance, and healthcare are also beginning to face automation risks.

Secondly, the speed of job replacement is accelerating. Deloitte's forecast shows that by 2026, 92 million jobs worldwide will be replaced by artificial intelligence, while 17 million new jobs will be created. This rapid replacement places higher demands on workers' skill updates.

Third, the pattern of income distribution may change. The application of artificial intelligence technology could further widen the gap between capital income and labor income, especially having a more significant impact on medium-skilled workers. This trend may exacerbate the existing issue of income inequality.

Regional Economic Warning Signals

Economic data from Texas provides important warning signals. In October 2025, the state's service sector revenue index fell to -6.4, the lowest level since July 2020. The employment index was -5.8, and the business activity index was -9.4, both indicating a significant contraction trend.

The performance of the retail sector is more severe, with the sales index dropping to -23.5 and the employment index falling to -5.3. These data align with the overall economic trend in the United States, where national retail sales increased by 0.6% month-on-month in August, but the core sales growth rate was only 1.5%, reflecting insufficient consumer momentum.

Labor market indicators also show signs of pressure. The consumer confidence index fell to 94.6, and the labor disparity index rose to 9.4%. These changes are temporally correlated with the promotion and application of artificial intelligence technology, suggesting that technological changes may be impacting the job market through multiple channels.

Risk Assessment Framework

From a macroeconomic perspective, the employment risks brought by artificial intelligence are mainly reflected in the following aspects:

In the capital markets, the median price-to-earnings ratio of the top 10 artificial intelligence companies in the S&P 500 index has reached 32 times, significantly higher than the market average. This valuation disparity may reflect the market's overly optimistic expectations for artificial intelligence earnings, and if actual earnings fall short of these expectations, it could trigger a market adjustment.

The relationship between productivity and employment is also worth noting. In the second quarter of 2025, non-farm productivity in the United States grew by 3.3%, while unit labor costs only increased by 1.0%. If this gap continues to widen, it may indicate that the gains from productivity improvements are not being fully translated into workers' income, which in turn could affect overall demand.

From a historical comparison, the current situation has certain similarities with the 1930s. At that time, technological advancements also led to a significant increase in productivity, but due to issues such as insufficient demand and income distribution, it ultimately exacerbated employment pressure. This historical experience reminds us to comprehensively assess the employment effects of artificial intelligence.

Policy Response Plan

Based on historical experience and current analysis, effective policy responses should include the following elements:

Reforming the education system is a long-term foundation. It is necessary to focus on strengthening the cultivation of data literacy, analytical skills, and innovative thinking, and to establish a curriculum system and vocational training mechanisms that are compatible with the era of artificial intelligence. Special attention should be paid to the construction of a lifelong learning system to help workers cope with the frequent demand for skill updates.

The improvement of the social security system is crucial. This includes expanding the coverage of unemployment insurance, establishing vocational transition assistance programs, and exploring social security systems that adapt to new employment forms. During the period of technological transformation, a well-developed social safety net can effectively reduce transition costs.

Industrial policies need to play a guiding role. Encouragement should be given to the deep integration of artificial intelligence with traditional industries and support for the development of emerging industries, compensating for job losses caused by technological replacement by creating new employment opportunities. At the same time, attention should be paid to regional coordinated development to avoid excessive concentration of employment impacts in specific areas.

Conclusion and Outlook

Artificial intelligence technology is triggering a new round of employment structure adjustment. Historical experience shows that technological unemployment has cyclical and structural characteristics, and its depth and duration of impact depend on the pace of technological advancement, labor market flexibility, and the effectiveness of policy interventions.

Amazon's layoff decision reflects the adaptive adjustments of companies at the enterprise level in response to technological changes. From a macro perspective, this adjustment is a necessary process for improving resource allocation efficiency, but it also brings friction to the job market. A successful transformation requires collaborative efforts from the government, businesses, and educational institutions to reduce the costs of transformation through institutional innovation and achieve social sharing of technological dividends.

Future research should focus on the heterogeneity of the impact of artificial intelligence on different skill groups and the adaptability of regional labor markets. At the same time, a more comprehensive data monitoring system needs to be established to timely assess the employment effects of technological changes and provide a scientific basis for policy-making.

Ultimately, the employment issue in the era of artificial intelligence is not only related to economic development but also concerns social stability and people's well-being. Only through systematic policy design and the joint efforts of the whole society can we achieve the coordinated development of technological progress and employment stability, promoting society towards a more inclusive and sustainable direction.