The artificial intelligence race between China and the United States has reached a critical stage. By early 2025, DeepSeek’s chatbot matched the performance of OpenAI’s ChatGPT. However, regardless of who wins the race, the more crucial question globally is: How will AI-driven productivity boosts impact labor markets as companies automate key business tasks on a large scale?

Absent careful planning, rapid adoption of AI risks widespread job displacement. Policymakers and AI developers therefore face an urgent challenge: Orchestrating transitions that augment rather than abruptly replace human labor. Despite substantial AI investments in both China and the US, reskilling programs, transition funding, or institutional safeguards to support workforces remain limited. AI investment in both countries is surpassing policies supporting labor forces. Indeed, both countries are intensifying tech competition while simultaneously pushing AI into everyday economic life.

China has been especially explicit. Shenzhen’s municipal government has unveiled a five-year plan to embed AI in every household and across the city’s business ecosystem by 2030.

Financial markets are responding. A January rally in Hong Kong was driven in part by Chinese AI stocks, including the initial public offering of Zhipu, and MiniMax, another leading Chinese AI firm, surged as much as 78 percent on its first day of trading on Jan 9.

US markets tell a parallel story, with the “magnificent seven” stocks (Nvidia, Apple, Alphabet, Microsoft, Amazon, Meta Platforms, and Tesla) posting an average return of 27.5 percent in 2025. Alphabet and Nvidia, in particular, dramatically outpaced the S&P 500, returning 65 percent and 40 percent, respectively. Capital is betting heavily that AI will remain the dominant growth engine.

On the positive side, AI-driven productivity gains are real. Robotaxis by Baidu’s Apollo Go and Alphabet’s Waymo are gaining momentum. Autonomous vehicles are rapidly advancing in China and the US. Around the world, firms are deploying autonomous agents for coding, customer support, and data analysis to increase throughput, accuracy, and decision speed. Agentic AI systems can even coordinate multistep workflows, exemplified by Microsoft Copilot’s ability to operate across email, calendars, and documents.

In finance and professional services, recent research shows that AI can already perform a significant share of tasks traditionally done by US workers. This capability could save up to $1.2 trillion in wages, demonstrating substantial productivity gains. Corporations that integrate AI-based solutions are therefore poised to increase revenues by both increasing labor productivity and reducing labor-related costs.

The race to deploy AI should not be a race to hollow out the social contract. By treating productivity as not only a private but also a public good, we can build an AI-enabled economy that is both efficient and humane

While many firms are hiring experts to develop AI agents to boost productivity, the addition of these technical staff members is overshadowed by the redundancy of traditional workforces. Therefore, AI-driven productivity gains can result in significant human job losses. AI expert Geoffrey Hinton predicts that AI will replace many jobs worldwide in 2026. For instance, Ted Egan, San Francisco’s chief economist, has noted that AI-driven worker displacement is a key factor contributing to the recent trend of job losses in the city’s technology sector. Khan Academy CEO Sal Khan has raised concerns about autonomous taxis and AI agents displacing large shares of call center work in the Philippines, bringing global services into the automation frontier. McKinsey estimates that by 2030 up to 30 percent of current work hours could be automated, catalyzing as many as 12 million occupational transitions, with a disproportionate impact on lower-wage workers.

To advance AI development while minimizing abrupt workforce displacement, we propose the following policy interventions.

Firms that deploy AI to replace human labor have a moral and social responsibility to sustain meaningful human employment. Governments must ensure that AI adoption serves the public interest. When firms fail to meet this responsibility voluntarily, policymakers should intervene through regulation or legislation to protect workers and communities from structural economic harm.

Next, as AI-enabled systems increasingly replace human roles in supply chains, such as driverless trucks and in knowledge-intensive fields like accounting, finance, law, marketing, and software engineering, there remains steady demand for human-intensive work in construction, repair, and maintenance. Moreover, in healthcare, nursing, therapy, and caregiving, especially in aging societies, there is a persistent preference for human provision rooted in empathy and compassion.

Indeed, firms can also share AI-driven productivity gains by reducing working hours rather than cutting headcount. If AI enables equivalent output with fewer labor hours, companies can preserve earnings, improve work-life balance, and retain institutional knowledge by maintaining full compensation while shortening the traditional workweek.

Educational institutions, such as vocational schools, should move quickly to develop curricula that train displaced workers for human-intensive roles where demand is rising. Companies that capture AI-driven profits should allocate a proportion of these profits to training programs for redundant workers.

Tax policy should be modernized. An AI tax akin to the robot tax proposed by Bill Gates could fund universal basic income and reskilling, cushioning transitions while preserving incentives to deploy AI where it truly adds value.

AI’s job impacts are not destiny but design. We can capture productivity gains while minimizing social costs if policy and corporate practice jointly fund transitions, reward continuous learning, and reform tax systems to support the most vulnerable. More effective responses to AI-driven labor disruption, therefore, lie in strategic combinations of skills first hiring, modernized social safety nets, and enforceable employer accountability for worker redeployment and transition. The race to deploy AI should not be a race to hollow out the social contract. By treating productivity as not only a private but also a public good, we can build an AI-enabled economy that is both efficient and humane.

Mark Pirie is a trauma psychologist who engages in independent research on all forms of trauma. 

Christopher Tang is a distinguished research professor at the UCLA Anderson School of Management.

The views do not necessarily reflect those of China Daily.