Who will be thrown overboard

**Who Will Be Thrown Overboard**

The global labor market is approaching a structural shift that differs fundamentally from past technological changes. For two centuries, mechanization mainly reduced the need for physical labor while simultaneously creating jobs in industry, trade, and the service sector. Current automation is gradually affecting intellectual and administrative functions long considered relatively safe from technological replacement.

Among the professions most exposed to change are cashiers, contact center operators, dispatchers, accountants, bank clerks, and administrative staff. According to a World Economic Forum forecast published in 2025, global structural changes by 2030 could lead to the disappearance of 92 million jobs and the creation of 170 million new ones. However, a positive net balance does not mean that workers who lose their jobs will be able to fill the emerging vacancies. New professions require different skills, often imply higher qualifications, and are concentrated in a limited number of economic hubs.

For Kazakhstan, Uzbekistan, Kyrgyzstan, and Tajikistan, this issue is especially significant due to a relatively young population structure and a large share of mid-skilled workers. The main threat lies not so much in an absolute decline in employment as in the disruption of traditional mechanisms for transitioning between professions. An economy may simultaneously experience a shortage of qualified specialists and face a growing number of people whose professional skills are no longer in demand.

Historical experience shows that technological progress is by no means always accompanied by a rapid recovery in employment. During the Industrial Revolution in Great Britain, the mechanization of textile production significantly reduced the need for certain categories of artisans. Industry later created many jobs, but the transition was accompanied by declining incomes for specific professional groups and prolonged social adjustment. Similar processes occurred in the United States in the second half of the twentieth century, when manufacturing automation and off-shoring reshaped the economic structure of industrial regions. Studies by economists Daron Acemoglu and Pascual Restrepo showed that the introduction of industrial robots had a negative impact on employment and wages in the most affected local labor markets. At the same time, overall economic growth did not automatically restore the previous level of worker welfare.

A significant difference of the current stage lies in the speed at which software solutions spread. While building a robotic factory requires capital investments and several years of preparation, deploying an intelligent document processing system can take considerably less time. Therefore, the time gap between the disappearance of old functions and the emergence of new professions is capable of widening. It is precisely this gap that becomes an independent economic factor influencing consumption, social mobility, and the sustainability of public finances.

The most vulnerable segment is mass office employment, which took shape thanks to the expansion of the banking sector, trade, telecommunications, and public administration. Traditionally, such jobs provided a relatively stable income without the need for extensive technical education. However, a significant portion of the operations performed relies on repetitive actions, standardized rules, and information processing. Modern systems are capable of automatically classifying customer requests, verifying documents, generating accounting entries, and routing transport assignments. In its studies on the impact of generative artificial intelligence, the International Labour Organization singled out clerical and administrative occupations as categories with a high degree of technological exposure. At the same time, vulnerability to automation does not equal the total disappearance of a profession, as many workflows require human oversight. Nevertheless, even partial substitution of tasks can significantly reduce staffing requirements. If ten operators used to serve a certain flow of requests, and a new system allows three employees to handle it, an enterprise gets the opportunity to downsize without changing the volume of services. The mass adoption of such solutions gradually turns individual employer savings into a structural problem for the national labor market.

For Central Asia, the consequences may prove particularly complex due to demographic differences and varied economic structures. The population of Uzbekistan already exceeds 39 million people, with numerous young age groups increasing the labor supply every year. Kazakhstan has a more diversified economy, yet a significant share of urban employment is also concentrated in trade, finance, education, and administrative services. Kyrgyzstan and Tajikistan stand out for the heavy reliance of households on labor migration and remittances. According to World Bank estimates, in certain years remittances to Tajikistan accounted for over 40% of gross domestic product, indicating the economy's high sensitivity to external demand for workers.

Automation will affect not only domestic labor markets, but also host countries, including Russia. The spread of automated warehouses, electronic document management, and digital services is capable of altering the structure of demand for migrants. At the same time, a reduction in job openings in some sectors may coincide with persistent labor shortages in construction, utilities, and elderly care. Consequently, migration will not disappear, but its professional composition may change noticeably.

A separate issue is created by the mismatch between the education system and the real dynamics of demand for skills. Universities and colleges in Central Asia graduate specialists in accounting, finance, management, and public administration every year. A significant share of educational programs is oriented toward a professional structure formed prior to the widespread adoption of intelligent systems. Meanwhile, automation is capable of reducing the number of entry-level positions through which young professionals traditionally gained practical experience. In the past, an economics graduate could start a career by processing primary documentation and gradually move on to analytical duties. If basic operations are handled by software, it is more profitable for an employer to hire a small number of experienced specialists. A contradiction arises between growing organizational productivity and shrinking opportunities for youth professional development. It may become especially noticeable in major cities, where higher education is perceived as the main engine of social mobility. As a result, a formal increase in the number of credentialed specialists will not necessarily be matched by growth in demand for them. Educational expansion accompanied by insufficient restructuring of course content can exacerbate skill imbalances.

The economic consequences of automation extend far beyond direct job losses. Wages represent both production costs for enterprises and a source of consumer demand. When a company reduces staff, it increases operational efficiency; however, laid-off employees cut back their spending on goods and services. If such processes occur simultaneously in banking, trade, and administrative services, the negative impact on domestic demand can become systemic. For Central Asian economies, where household consumption plays a significant role, this mechanism is particularly critical.

An additional factor is the distribution of income gains from productivity growth between capital owners and wage earners. In 2024, the International Monetary Fund estimated the share of global employment potentially affected by artificial intelligence at approximately 40%. However, the degree of impact varies noticeably depending on an economy's level of development and the nature of the tasks performed. High-performing companies are able to boost profits without a comparable expansion of headcount. If revenues concentrate among a limited circle of technology owners, overall economic growth may be accompanied by widening wealth inequality.

Another consequence is a shift in the spatial organization of employment. Modern digital sectors exhibit a pronounced tendency to cluster in large cities equipped with universities, investment capital, and developed infrastructure. In Kazakhstan, Almaty and Astana primarily hold these advantages, whereas in Uzbekistan a significant share of tech activity is concentrated in Tashkent. For smaller towns, the automation of administrative functions can mean losing a limited number of relatively well-paying jobs. Unlike capital hubs, regional economies often lack a sufficient number of alternative employers. The movement of workers to megacities places added strain on housing markets, transport, and municipal infrastructure. Simultaneously, peripheral territories lose solvent populations, exacerbating conditions for local trade and services. A cumulative mechanism of spatial inequality emerges, whereby technological advantages concentrate where advanced economic institutions already exist. Historical examples from industrial regions in Europe demonstrate that the consequences of structural job loss can persist for decades. Therefore, the geographic distribution of new occupations is no less important than their total count.

The widespread belief that displaced employees will automatically transition into tech industries does not sufficiently account for professional constraints. A machine learning specialist's job demands mathematical training, programming, and practical experience that cannot be acquired through short-term courses alone. For a middle-aged accountant or a contact center operator, moving into such a profession may require several years of study. Furthermore, not all new jobs are directly tied to technology development. Demand growth is possible in healthcare, maintenance, energy, construction, and social services. However, these sectors impose their own requirements regarding physical capability, professional certification, and geographic mobility. In addition, some types of employment offer relatively low pay, making the transition economically unequal. A bank employee who loses an administrative position can formally find work in delivery services, but such a career change is often accompanied by a decline in income and social security. Consequently, a statistical recovery in employment may conceal a deterioration in job quality. This is precisely why the number of vacancies is not a sufficient indicator of an economy's successful technological adaptation.

Substantial changes also await the public finance system. In most Central Asian countries, mandatory social contributions and income taxes are closely linked to formal employment. A drop in the number of workers alongside rising enterprise productivity is capable of altering the ratio between payroll and value added created. If the tax system relies heavily on labor income, the state risks capturing a smaller share of the economic benefits of automation. Simultaneously, the need to fund professional retraining, social safety nets, and employment assistance programs increases. For states with limited fiscal capacity, a conflict arises between the necessity to incentivize tech investments and maintaining the stability of social obligations. Added complexity comes from a large informal sector, where labor relations frequently fail to provide adequate insurance coverage. Workers moving from formal jobs to irregular services may formally remain economically active while losing pension savings and social guarantees. Thus, technological modernization is capable of transforming not only the production structure, but also the financial foundation of existing social institutions.

At the same time, automation does not necessarily mean a long-term decline in the total number of jobs. Productivity gains lower the cost of individual services, broaden product accessibility, and potentially generate additional demand. The history of automated teller machine (ATM) adoption illustrates the ambiguity of such processes: lowering the cost of banking services enabled branch networks to expand and employee roles to be reassigned under certain conditions. However, similar compensating mechanisms are not guaranteed with the deployment of generative artificial intelligence. Digital technologies possess high scalability, as a single software system can serve numerous enterprises without a proportional increase in personnel. This distinguishes them from infrastructure technologies, the rollout of which is typically accompanied by substantial demand for construction and operational labor.

For Central Asia, dependence on imported digital solutions carries added weight. If core software platforms are developed outside the region, a significant portion of the value added created may concentrate among foreign vendors. In that case, national economies gain efficiency boosts, but not necessarily a corresponding increase in their own tech employment. It is therefore vital to distinguish between adopting new technologies and participating in the formation of associated production chains.

In the long run, the most likely outcome is not the disappearance of labor as an economic category, but a shift in its internal hierarchy. Occupations based on standardized information processing are gradually losing their former stability, while the relative value of complex technical, managerial, and interpersonal competencies rises. Simultaneously, substantial demand persists for physical labor that is technically harder to automate in unpredictable environments. A risk of employment polarization arises, where high-paying specialized roles and relatively low-paying service jobs expand while the intermediate professional layer contracts. For Central Asia, this process overlaps with demographic growth, regional inequality, and the ongoing reliance of individual states on labor migration. Even with positive dynamics in gross domestic product, the social consequences may prove contradictory if productivity grows faster than incomes for a significant portion of the population. The main bottleneck for economic development will then be not a lack of technology, but an inability to translate technological advances into broad public participation in wealth creation and distribution. Mass automation effectively tests the capacity of existing economic institutions to ensure worker transitions between professional paradigms without prolonged exclusion from productive activity. This constitutes the fundamental contradiction of the upcoming transformation: society can produce more goods and services with less labor input, while simultaneously confronting growing economic insecurity among those whose professions once underpinned the stability of the middle class.