The employment substitution risk of artificial intelligence (AI) has entered the policy perspective. The General Office of the State Council recently issued the "Opinions on Adapting to the Development of Artificial Intelligence Technology and Supporting the Construction of an Employment Friendly Development Mode" (hereinafter referred to as the "Opinions"), which makes systematic arrangements from creating employment opportunities, slowing down employment substitution, strengthening skills training to improving social security. On November 3rd, the National Development and Reform Commission proposed at a press conference to strengthen monitoring and early warning, enhance the connection between the social security system and the reform of employment structure and work methods, and ensure a comprehensive and inclusive social security.

These deployments indicate that the impact of AI on employment is no longer a distant hypothesis, but a real issue that needs to be addressed. Undoubtedly, the new scenarios and applications brought by AI have given rise to a group of new professions and positions, but the impact it brings is also gradually becoming apparent. Especially, due to the development of AI technology and the consideration of cost-benefit factors by enterprises, various applications may not necessarily enter every aspect of economic life according to our expected path and pace.

For example, we hope that AI can be prioritized for promotion and application in positions with labor shortages and high-risk environments - there have been many successful cases and they are still being promoted on a larger scale. But under the pressure of cost reduction and efficiency improvement, companies are likely to focus more on labor-intensive and basic positions. In many highly competitive industries with limited profit margins, this is also a story happening.

To what extent can we influence their entry path selection and pace, especially in labor-intensive industries? How can we promote AI applications that prioritize people and strive for excellence while respecting market laws? This is a mandatory question. It is precisely because of this that the Opinion proposes to "mitigate the impact of artificial intelligence applications on employment substitution" and requires a "comprehensive evaluation of the impact of artificial intelligence applications on employment, guiding innovation resources to tilt towards areas with great potential for job creation".

In real life, the public sector, state-owned units, and private enterprises may face different constraints, and the use of AI is not entirely dependent on efficiency and cost. For example, personnel adjustments may be constrained by factors such as staffing, organizational systems, and public service responsibilities. Faced with similar technological advancements, there may be significant differences in the employment risks borne by different workers.

In other words, the "susceptibility" of different departments, industries, and types of enterprises to technological shocks varies. This difference truly reflects the economic structure, which in turn is reflected in the effectiveness of policy guarantees. For example, whether job changes can be gradually digested within existing organizational and personnel management systems, and whether policy requirements can be more easily translated into specific tasks and assessments, all of which may affect the feelings and immediate interests of practitioners when shocks arise.

If we only consider the influence of market factors, it can be said that the more competitive the fields are, the more motivated they are to promote the application of AI, and practitioners in these fields are also more likely to encounter employment substitution risks. Before the broader impact arrives, this group of workers may have already endured the shock. The problem is that the decisions made by individual enterprises based on their own reality may have market rationality, but the resulting comprehensive impact may be transformed into social issues that cannot be ignored.

This tells us that at a time when AI is penetrating thousands of industries, stabilizing employment not only requires stabilizing the number of jobs, but also ensuring that those who have lost their jobs due to technological progress can still have security and opportunities for further development.

The discussion on which professions and occupations are more likely to be replaced in the AI era has been ongoing for the past two years. Many times, rather than needing a definite answer, it is more accurate to say that both individuals and society hope to be prepared for potential risks. Therefore, the Opinion encourages enterprises to explore the system of unemployment buffer period, inform affected workers in advance during the large-scale promotion and application of artificial intelligence, and provide support for transition and employment.

These measures are targeted, but it is also important to note that the solid implementation of encouraging arrangements is often related to the business situation and willingness of the enterprise. This also reminds us that employment policies should not only respect the laws of technological progress and enterprises' autonomous adjustment of employment, but also share the costs that enterprises find difficult to bear alone during the transformation process through unemployment protection, public employment services, and vocational training. Policy support should follow workers more closely, rather than relying on their original units and positions.