On September 10, the 2026 Inclusion·Bund Financial Conference opened with the theme of "Creating a New AI Economy." Focusing on this theme, Philippe Aghion, the Nobel laureate in economics and French economist, delivered a speech titled "Creative Destruction" and the AI Revolution. He believes that AI has the potential to increase productivity growth by an additional 1.08 percentage points annually over the next decade, which could have an impact greater than the IT revolution. However, whether this technological potential can be transformed into sustained growth depends on whether competition policies, education systems, and labor market institutions can keep up simultaneously.
Aghion first analyzed the economic impact AI might bring from the perspective of productivity. His research with colleagues shows that considering only the automation of tasks in the production of goods and services, AI has the potential to increase productivity growth by an additional 0.68 percentage points annually over the next ten years, which is on par with the productivity gains estimated from the experience of the IT revolution.
"0.68 percentage points is just a minimum," he said. If we also consider the role of AI in promoting new ideas, Aghion believes that productivity growth could increase by at least another 0.4 percentage points annually, leading to a potential total increase of about 1.08 percentage points.
AI can push productivity higher, but technological breakthroughs do not automatically lead to economic growth.
However, he also warned that technological breakthroughs do not automatically translate into sustained growth. Aghion cited the example of the U.S. IT revolution. Between 1996 and 2005, the information technology revolution drove a significant increase in total factor productivity in the U.S., and a group of "superstar companies" rapidly grew; meanwhile, industry concentration continued to rise, and the entry rate of new companies began to decline around 2000, followed by a noticeable slowdown in productivity growth.
"This is a core contradiction in the theory of creative destruction: innovation requires that successful individuals receive sufficient rewards to sustain innovation; however, once these innovators succeed, they may use their existing advantages to limit the space for new entrants and innovation."
Aghion believes that this issue should also be addressed in the AI era. Currently, the upstream of the AI value chain has already shown a high level of concentration, with the cloud computing market dominated by large companies such as Amazon, Google, and Microsoft, and the graphics processing unit market is also highly concentrated. "AI indeed has great growth potential, but whether it can be fully realized also depends on how well we handle competition issues," he said.
When discussing the experiences of different economies moving from catching up to frontier innovation, Aghion said that productivity growth can be achieved either through imitation and absorption of existing technologies or through frontier innovation. As an economy gets closer to the technological frontier, the importance of the innovation environment will further increase.
Regarding China, he said that, based on indicators such as the proportion of high-tech patents, China is currently performing well in frontier innovation. "China is doing very well in frontier innovation," he added. At the same time, he pointed out that Europe is not only lagging behind the U.S., but also falling behind China in frontier innovation.
Education and the labor market must learn to "catch people" in response to the impact of the AI new economy.
The other side of the AI new economy is what impact it will have on ordinary people after productivity increases.
Aghion believes that AI will take over some tasks previously performed by humans, resulting in fewer jobs in some areas; however, companies adopting AI will see improved productivity and competitiveness, leading to increased demand and potentially more employment. Additionally, AI's promotion of new ideas will also create new products, companies, and job positions.
"AI will certainly replace some human labor, at least some of the tasks," Aghion said. "But on the other hand, AI will also create new job positions by increasing productivity and inspiring new ideas."
He stated that responding to changes brought by AI requires a good education system, and "China has the conditions for this." Education is not only about teaching specific knowledge but more importantly about helping people learn how to learn, so they can continuously adapt to technological and work-related changes. He pointed out, "We have never needed more to enhance the value of manual skills and soft skills."
In terms of education, Aghion also specifically discussed the boundaries of AI entering the classroom. He suggested reserving part of the learning process that does not use AI at all, allowing students to independently read, calculate, and think; the AI used by students should be carefully selected and regulated, and the risks associated with AI should be made known. "If AI is left unchecked, we may see an entire generation unable to calculate independently or focus on reading, which would be very harmful," he emphasized.
Regarding the labor market, he used Denmark's "flexible security" system as an example, suggesting that during job transitions, unemployment benefits, retraining, and support for reemployment can help workers more smoothly transition into new jobs.
Aghion concluded that to release AI's growth potential, appropriate competition policies are needed; to release AI's potential to create jobs, a good education system and labor market protection are essential.
From productivity growth, frontier innovation to changes in employment, Aghion's speech provides an economic perspective for understanding the AI new economy: the new growth brought by AI not only depends on what technology can do, but also on whether new innovations can continue to emerge, and whether more businesses and workers can participate in and benefit from this process.
