Yicai Editorial: Establish First, Break Later—Creating Controllable Boundaries for Breakthroughs in AI Applications

Main text

Seeking all controllable certainties is becoming an important demand in artificial intelligence applications.

Recently, the General Office of the State Council issued the Opinions on Adapting to the Development of Artificial Intelligence Technology and Helping to Build an Employment-Friendly Development Model, clearly deploying 18 major tasks in six areas, including strengthening the employment-creation effect of artificial intelligence, building an employment-friendly path for AI upgrading, and improving workers' digital literacy and employability, aiming to build an employment-friendly development model and buffer the irreversible reshaping and restructuring impact of artificial intelligence on the economy and society.

Artificial intelligence is undoubtedly bringing enormous changes to the economic and social ecosystem, and this high degree of uncertainty about the future allows panic to preemptively influence people's judgments. In the history of human technological progress, similar panics have played out countless times, because technological revolutions change each person's initial configuration in the manner of destructive creation, break the stability people possess, and trigger people's loss-aversion effect.

Building an employment-friendly AI development model can be seen as exploring an emotional intermediary mechanism that establishes a connection between the present and the future. Its essence lies in guiding a reduction of shocks, allowing AI applications to establish connections with people in a smoother way, and setting the impact of AI on the economic and social ecosystem within a range people can bear as much as possible.

At the same time, building an employment-friendly AI development model reveals the inherent consistency between artificial intelligence and stable employment, namely that AI applications will bring more gains to employment, and replacement will be more structural. Technological development has typical characteristics of punctuated equilibrium; the real application of new technologies in the economy and society is a journey of integration, and only evolution based on cost minimization can reduce confrontation in the process of replacing the old with the new as much as possible.

Therefore, the evolution of artificial intelligence in the economic and social ecosystem follows a logic of mutual gain. Take the software industry, which AI has hit first in recent years, as an example: people have not seen large-scale replacement of computer software engineering jobs. On the contrary, programmers are the group that has most actively embraced AI, mastered AI, and enhanced work efficiency, and the software industry will also be the first industry to enjoy AI's gains.

Lower marginal costs, higher marginal efficiency, and faster and more sensitive response speeds are the tangible presentations AI currently brings to people. Clearly, the Opinions' explicit strengthening of AI's employment-creation effect truly conveys the positive role of AI in helping small and medium-sized enterprises stabilize and expand employment. In particular, as AI is moving from large language models into the development stage of physical world models, AI needs to become familiar with, understand, and apply physical rules and achieve a leap from language interaction to behavioral interaction, which requires a large number of small and medium-sized enterprises and workers to serve as guides for physical AI and help physical AI understand the entire physical world. This dooms AI's gain-enhancing function for the vast number of workers; people's panic about AI replacing jobs is more about how to get along with AI.

Building an employment-friendly path for AI upgrading is not a conservative guide for technological application, but should be understood as revealing the essential connotation of the integration of AI with the existing economic and social ecosystem. Human-computer interaction is the only way for AI to integrate into the economic and social ecosystem.

What people need to actively move away from now is the mindset of confrontation with AI, and they should cautiously avoid placing themselves in a replacement position of competing on the same stage with AI. Only in this way can they face squarely the gains that AI and other technological advances bring to them.

This leads to a new cognitive change: building employment-friendly AI starts a two-way journey. On the one hand, AI aims to liberate people from standardized workflows; on the other hand, it is necessary to continuously improve workers' digital literacy and employability, guide workers to better understand and use AI, enable them to grow into AI-innovative talent, and make clear that future workers' competitiveness will center on how to better use and cooperate with AI in work.

Therefore, it is not difficult to find that the Opinions convey a belief in AI development of establishing the new before abolishing the old. This establishing the new before abolishing the old is, on the one hand, a strategic arrangement for restructuring and reshaping the economic and social ecosystem, namely explicitly requiring the tapping of AI employment potential in traditional fields and mitigating the employment-replacement impact of AI applications, aiming to reduce AI's direct shock to the job market and provide more space and time for people to correctly understand AI.

On the other hand, people's understanding of AI requires establishing the new before abolishing the old in terms of beliefs and concepts. People's current panic about AI replacing jobs is not mainly based on real evidence, but on speculation about possibilities. This is essentially a question of how to understand and get along with AI. Only by moving beyond AI panic and truly actively embracing AI can AI possibly become a gain rather than a deprivation, and can humans and AI become an efficiency combination and produce a mutually reinforcing effect.

Building an employment-friendly path for AI technology development is to create and build a journey of integration in which humans and AI mutually gain, and to create controllable boundaries for breaking through in AI applications. Only by establishing this cognitively will people no longer be shrouded by the loss-aversion effect when facing AI.

(Source: Yicai) Middle of the main text Embedded advertisement