1. Research background
Artificial Intelligence (AI), as the core of the fourth industrial revolution, is an important technical force to promote the country's economic and social development. The Outline of the 14th Five-Year Plan for National Economic and Social Development of the People's Republic of China and the Long-Range Goals for 2035 clearly states that it is necessary to target frontier fields such as artificial intelligence and "implement a number of forward-looking and strategic major national science and technology projects". As an emerging technology in the era of digital economy, artificial intelligence has gradually penetrated into the field of unconventional work tasks such as copywriting and analysis and decision-making, and is replacing human labor with unprecedented breadth and depth, causing deep concerns from all walks of life that artificial intelligence will lead to unemployment of workers.
Theoretically, the substitution effect and recovery effect of artificial intelligence on employment have been mainly discussed, while more micro mechanisms need to be further explored to fully understand the internal mechanism and multiple paths of artificial intelligence affecting employment. The relevant empirical literature mainly uses the number of industrial robots to measure the level of artificial intelligence and investigate its impact on the employment of manufacturing enterprises, and it is yet necessary to construct indicators that can more accurately measure the level of artificial intelligence technology of enterprises and expand the analysis perspective to include the service industry. Therefore, the research on the impact of artificial intelligence on employment urgently needs to expand the micro-mechanism analysis from the theory, enhance the accuracy of measurement empirically and expand the research sample, so as to provide academic support for enterprises to implement artificial intelligence development strategies and governments to formulate artificial intelligence policies.
Second, the study found
Yin Zhifeng, Cao Aijia, Guo Jiabao and Guo Dongmei's paper "Research on the Employment Effect of Artificial Intelligence Based on Patent Data: Micro Evidence from Zhongguancun Enterprises" published in the 5th issue of China Industrial Economics in 2023 provides a theoretical analysis of the substitution effect and recovery effect of artificial intelligence, and conducts a malleable discussion based on the characteristics of China's industry, labor market and artificial intelligence technology. In terms of the theoretical mechanism of artificial intelligence affecting enterprise employment, this paper proposes a new mechanism of "market expansion effect" from the perspective of enterprise product market share expansion. In terms of empirical evidence, this paper combines the Chinese Gonggong Intelligent Patent Database and Zhongguancun Enterprise Survey Database from 2009 to 2018, uses the number of enterprise artificial intelligence patent applications to measure the level of artificial intelligence technology, and examines the impact of artificial intelligence on enterprise employment and its micro-mechanism. The main findings are as follows:
First, artificial intelligence has significantly increased the employment level of enterprises. At present, the recovery effect of Chinese labor intelligence promotes the formation of "new employment forms", which is dominant compared with the substitution effect, and is generally manifested in improving the employment level of enterprises. China's artificial intelligence is developing rapidly at the application level, around the development of artificial intelligence, supporting front-end development, mid-end operation and back-end maintenance positions have emerged, and enterprises' in-depth exploration of its application scenarios is also giving birth to various new models and new formats, creating more labor demand. In contrast, the continuous upgrading of China's employment structure also provides more space for the recovery effect of artificial intelligence and compensates for the employment loss caused by the substitution effect.
Second, artificial intelligence has a "market expansion effect" by increasing the market share of enterprise products and promoting the expansion of employment scale. On the one hand, artificial intelligence enhances the innovation ability of enterprises, helps enterprises effectively capture market signals, predict market fluctuations, and quickly respond to market demand to seize market share. On the other hand, artificial intelligence reduces the production cost of enterprises, which is conducive to reducing the relative price of products, and the rise in product demand and the increase in corporate profits further stimulate enterprises to expand reproduction. With the continuous expansion of the market size of enterprises, those departments within enterprises that are not easily replaced by artificial intelligence will recruit more labor, driving the overall employment growth of enterprises.
Third, artificial intelligence affects the employment structure of enterprises from three dimensions: education, age and seniority. In terms of academic qualifications, artificial intelligence has significantly increased the proportion of employees with college degree or above in enterprises, and significantly reduced the proportion of employees with college degree in enterprises. In terms of age, artificial intelligence has the effect of promoting the "youthfulness" of employment. Artificial intelligence has significantly increased the proportion of employees aged 30-39 in enterprises, the proportion of employed persons aged 40-49 and the proportion of employed persons aged 50 and above significantly decreased. In terms of working years, artificial intelligence will significantly increase the proportion of employees with 3 years or less in enterprises, and significantly reduce the proportion of employees with 5 years and above.
Fourth, the job creation effect of AI differs significantly between different technical segments and enterprises with different characteristics. Compared with expert systems and robots, the employment promotion effects of computer vision, machine learning and natural language processing are more obvious. The effect of artificial intelligence in promoting employment decreases with the increase of enterprise scale and total exports. Compared with private enterprises, the employment promotion effect of AI is smaller in state-owned enterprises and collective enterprises.
3. Research enlightenment
First, the rapid development of artificial intelligence at this stage is not a "flood beast" for employment. Relevant government departments need to guide and support enterprises to apply artificial intelligence technology, and take measures such as setting up special AI innovation and development funds for enterprises, building public data platforms, and building leading AI enterprises in the industry, so as to alleviate the financing constraints faced by enterprises' AI research and development, lower the threshold for enterprises to build AI systems, and give play to the synergy effect of industry innovation.
Second, artificial intelligence improves the innovation ability of enterprises, reduces the production costs of enterprises, and then drives the expansion of enterprise market share, and ultimately leads to the improvement of enterprise employment level. In view of this, policymakers can consider combining the current situation of economic development and the level of supply and demand in the job market, accelerate the development of the artificial intelligence industry by further formulating artificial intelligence-related industrial policies, encourage enterprises to apply more artificial intelligence-related technologies, strengthen their innovation capabilities and product quality, and increase the overall employment level of enterprises while driving the expansion of enterprise market share.
Third, artificial intelligence will trigger changes in the employment structure, which will not only drive employment growth, but also lead to structural unemployment. For the "beneficiaries" under the impact of artificial intelligence, on the one hand, the government can consider improving the construction of disciplines, setting up corresponding majors according to the actual requirements of artificial intelligence development, cultivating high-level innovative talents, and increasing the supply of high-skilled labor. On the other hand, the government can strengthen on-the-job training, reduce the friction between the development of artificial intelligence and the mismatch of labor skills, so that practitioners can better adapt to the requirements of human-machine integration under the development of artificial intelligence, and maximize the complementary role of artificial intelligence and labor. For the "damaged", the government can consider providing unemployment insurance for skilled unemployment, family minimum income guarantee and other "bottom-up" measures, provide laid-off workers with vocational training, help them complete human capital upgrading, and adapt to the requirements of the new technology era to divert low-skilled and high-age workers to other industries.
Fourth, the impact of different subdivisions of AI technologies on employment has its own merits, and the employment promotion effect of AI is also different in different types of enterprises. On the one hand, policymakers can consider combining the development status of the artificial intelligence industry and market demand, focusing on exploring and developing application scenarios of computer vision, machine learning and natural language processing, and promoting the application of these technologies in enterprises. For expert systems and robots, it is necessary to focus on their negative impact on employment. On the other hand, the government needs to weigh the depth and breadth of the use of AI, further deepen the market-oriented reform of state-owned and collective enterprises, and focus on market sensitivity and personnel mobility to give full play to the employment promotion effect of AI.
Originally published in: China Industrial Economics