At a quarterly forum on China's macroeconomy held in Beijing on June 27, Cai Fang, a member of the Chinese Academy of Social Sciences, delivered a speech.

He projected that to achieve the goal of becoming a moderately developed country by 2035, China's economy will need to maintain an annual growth rate between 4.4% and 4.8% over the next five years. As the demographic dividend fades, the efficiency gains from traditional resource reallocation are diminishing. However, artificial intelligence offers opportunities for more granular resource allocation, holding the potential to significantly boost labor productivity and break the downward trajectory of potential growth rates.

He concluded that AI primarily enhances productivity through two pathways. The first is the "resource allocation effect," where data and algorithms directly achieve optimal allocation, shifting from traditional "trial and error" to "trial and success." The second is the "creative destruction effect," as AI is expected to accelerate market selection and the elimination of weaker players at an unprecedented scale.

Addressing concerns about whether AI might fall into a productivity paradox—where new technology is ubiquitous but productivity gains are not visible—Cai Fang pointed out this is essentially a manifestation of uneven development. While AI intensifies "creative destruction" among firms, it also manifests two major negative macroeconomic effects.

The first is the "denominator effect." As companies pursue efficiency by reducing labor input, the number of jobs destroyed may exceed the number created.

The second is the "numerator effect." If supply-side capacity increases substantially without a corresponding growth in purchasing power on the demand side, AI could further exacerbate the current imbalance of "strong supply versus weak demand" and the overall insufficiency of aggregate demand.

Cai Fang believes the key to resolving this paradox lies in treating people and people's livelihoods as independent variables. He proposed the following three core measures.

First, create an upgraded version of proactive employment policies. It is essential to enhance these policies to counter the potentially larger employment shocks that future general artificial intelligence might bring.

Second, prioritize investing in people as the main method of redistribution. China's Gini coefficient for per capita disposable income has long remained at a relatively high level of 0.465, urgently requiring redistribution to improve income distribution. This would boost purchasing power and address the "numerator effect." Compared to tax adjustments, investing in people through social spending is more effective in improving income distribution.

Third, address the challenges of caring for the elderly and children through a "care economy." Cai Fang proposed transforming unpaid labor, such as household chores and caregiving, into paid labor, thereby developing a "care economy." This would not only create a large number of jobs and stimulate GDP growth but also, by using AI to replace dirty and arduous tasks, allow the care industry to fully embody human warmth, effectively addressing social pain points related to childbirth, child-rearing, and elderly care.

In conclusion, Cai Fang stated that while AI holds immense potential to resolve economic contradictions, it may also amplify existing developmental imbalances. Only by investing in people and tangibly improving livelihoods and income distribution can society truly embrace artificial intelligence and achieve inclusive economic and social growth.