Does artificial intelligence funnel wealth into fewer hands? Yao Yang, Dean of the Shanghai University of Finance and Economics' Lingang Advanced Finance Institute and a visiting professor at Peking University, isn't convinced. In a recent discussion on AI anxiety, he acknowledged AI as a defining opportunity of our era but pushed back against the notion that it will inevitably worsen social divides.

“Many say AI will concentrate wealth and widen social gaps. I don't believe it,” Yao Yang stated. He argued that history shows wealth was most extremely concentrated not in industrial societies, but in the agricultural era. Back then, land was the primary means of production. Vast landowners sat at the top, while the landless majority—laborers, tenants, and wanderers—had little chance to change their fate through personal effort.

“In the farming age, if you owned vast fertile fields, you were a great landlord at the very top. Most ordinary people could only work for others; that was when wealth was truly extreme in its concentration,” he explained.

In his view, while the Industrial Revolution had a harsh early phase, deeper industrialization gave more ordinary people the chance to earn a living through education and skills, gradually pushing society toward greater equality. “Without land, how could you make a living? There was almost no way out. After the Industrial Revolution, raising your education and productive capacity meant you could earn income through your own abilities,” he said, noting that technological progress expanded the division of labor and opened more economic avenues for ordinary citizens.

As global tech giants ramp up AI investment and capital markets trade actively around compute infrastructure and large-model supply chains, the debate over AI returns, valuations, and industry concentration has intensified. Yao Yang argues that high barriers in large-model development don't necessarily mean a few companies will dominate forever. “Some say a few big-model firms will monopolize the world. That hasn't happened, and it's unlikely to,” he said.

He pointed out that the large-model sector remains fiercely competitive, with companies constantly chasing each other through new models and technical approaches. “If you have a large model, I can build one too. First, there's competition, and competition itself imposes limits.”

Yao Yang also highlighted how China's large-model progress—especially the rise of open-source and low-cost solutions—is reshaping the global AI competitive landscape. Open-source models lower the entry barrier for businesses and individuals while putting pressure on proprietary foreign models. “Some Chinese large models are open-source and cost less. China has caught up, so don't underestimate this competition,” he cautioned.

With global competition intensifying, Yao Yang dismissed attempts to sustain a technological edge by restricting Chinese models as unrealistic. “What does banning Chinese large models even mean? Banning Americans from using them?” He suggested that many have yet to recognize how fundamentally the global AI competition structure has already shifted.