AI can support rural income growth, but only with infrastructure and policy backing

Artificial intelligence (AI) is often portrayed as a technology that could widen income gaps by rewarding skilled workers, automating routine jobs and concentrating gains in cities, but new research suggests a different outcome is possible when governments actively direct AI toward rural productivity, digital infrastructure and inclusive development. The study based in China finds that place-based AI policies can reduce urban-rural income inequality when they are linked to agricultural innovation, public investment and stronger government attention to the fair distribution of technological gains.

The research paper, titled Can Artificial Intelligence Narrow the Urban–Rural Income Inequality? Evidence from a Quasi-Natural Experiment in China, was published in Sustainability. It analyzes panel data from 257 cities from 2012 to 2023 and treats China's National New Generation Artificial Intelligence Innovation and Development Pilot Zone policy as a quasi-natural experiment to assess whether AI promotion strategies can reduce income inequality between urban and rural residents.

AI's inequality impact depends on how it is governed

In many economies, AI is expected to favor cities because urban areas usually have better internet networks, stronger universities, more skilled workers, higher investment capacity and closer links to technology firms. Rural areas, by contrast, often face weaker digital infrastructure, lower access to advanced services and limited capacity to absorb new technologies.

That imbalance has raised concerns that AI could deepen existing social and regional inequalities. If AI adoption remains concentrated in finance, manufacturing, software, urban governance and high-end services, the gains may flow mainly to already developed regions. The result could be faster productivity growth in cities and a widening gap between urban and rural incomes.

The study challenges that assumption by showing that the distributional effect of AI is not fixed. AI can widen inequality when left to market forces alone, but it can also reduce inequality when policy channels its use toward rural development, agriculture, public services and infrastructure. The central lesson is that AI's social impact depends not only on the technology itself but on the institutional environment in which it is deployed.

China provides a useful example because it has pursued AI development through national strategies, local pilot zones and policy-backed experimentation. The AI Innovation and Development Pilot Zone program was rolled out in stages from 2019, with designated cities encouraged to test AI applications, governance models and technology commercialization. The study uses the staggered rollout of these pilot zones to compare cities that received the policy treatment with cities that did not.

The researchers find that the pilot zone policy reduced the urban-rural income inequality index by about 8.41 percent. This result is important because it points to a measurable equalizing effect from AI policy, rather than only theoretical claims about inclusive technology. The finding remained robust after a series of checks, including alternative measures of inequality, placebo tests, matching methods, an instrumental-variable approach and exclusion of the COVID-19 shock year.

The findings do not suggest that AI automatically benefits rural households. The policy appears to reduce inequality through two main channels. First, it promotes agricultural science and technology innovation, improving rural productivity and raising the income potential of farmers. Second, it increases government attention to AI, which can redirect policy resources toward rural applications and more balanced technology diffusion.

This difference is crucial for countries trying to align AI adoption with sustainable development goals. Many governments want to use AI to improve growth and competitiveness, but fewer have clear strategies to ensure the gains reach rural populations, informal workers and low-income households. The China-based evidence suggests that AI can support inclusive growth only when supported by targeted infrastructure, agricultural innovation, training and governance.

Agricultural innovation turns AI into a rural income tool

The study identifies agricultural science and technology innovation as one of the strongest channels through which AI can reduce the urban-rural income gap. In rural economies, income growth depends heavily on productivity. Farmers and rural enterprises need better tools to raise output, lower costs, reduce waste, access markets and respond to environmental risks. AI can help, but only when rural producers have the infrastructure and policy support needed to use it.

AI-backed agriculture can include precision farming, smart irrigation, machine-learning-based crop monitoring, automated disease diagnosis, drone-assisted field management and data-driven advisory systems. These tools can help farmers use water, fertilizer and labor more efficiently. They can also improve crop quality, reduce losses and help rural producers respond faster to weather, pests and price changes.

In the case of China, the pilot zone policy appears to have encouraged innovation in agricultural technologies, including patent activity and wider deployment of AI tools. This matters because rural productivity gains can directly affect income distribution. When agricultural output rises and farming becomes more efficient, rural households can capture greater value from their work. If technology also improves access to markets, farmers may reduce dependence on intermediaries and gain more bargaining power.

The study also points to transaction costs as a major issue. Many rural producers face high costs in finding buyers, checking prices, verifying product quality and securing contracts. AI-enabled platforms and digital traceability systems can reduce these frictions. Farmers can access market information more quickly, while consumers and buyers can track product quality more reliably. In turn, rural goods may command better prices and reach broader markets.

A similar logic applies beyond China. In developing economies with large rural populations, AI could help strengthen agricultural value chains by connecting farmers to logistics, finance, market forecasts, weather information and technical advice. But this requires more than mobile apps. Rural regions need broadband access, data systems, digital literacy, affordable devices, extension services and policies that prevent large platforms from capturing most of the gains.

The study also highlights the role of skill barriers. Traditional theories of technological change often warn that advanced tools raise the wage premium for highly educated workers. That can worsen inequality. But AI may also reduce some technical barriers when designed for practical rural use. Voice-based advisory tools, automated crop diagnosis and simplified decision-support systems can help farmers use advanced knowledge without needing specialist training.

This does not remove the need for human capital. Farmers and rural workers still need basic digital skills to use AI effectively. But the findings suggest that AI can reduce some knowledge gaps when paired with training, subsidies and rural support systems. This is especially relevant for countries where rural workers have limited access to formal education but could benefit from practical, accessible technology.

To sum up, AI becomes an equalizing tool when it is tied to rural productivity rather than only urban automation. If governments want to narrow spatial inequality, they must direct AI investment toward agriculture, rural logistics, rural finance, rural tourism, local services and small-scale producers. Without that direction, AI may remain concentrated in urban sectors and deepen the divide.

China's pilot zones offer lessons for inclusive AI policy

The second key mechanism is government attention. When governments make AI a policy priority, they can shape where investment goes, which industries receive support and whether rural communities are included. In China's pilot zones, official attention to AI appears to have encouraged stronger policy support, greater resource allocation and a more active push to spread AI benefits beyond core urban sectors.

This finding has a broader significance. AI governance is often discussed in terms of safety, privacy, ethics and regulation. Those issues are important, but the study shows that AI governance is also about distribution. Governments must decide whether AI policy will mostly support elite industries and large firms, or whether it will also strengthen rural development, social services and regional equity.

The Chinese example shows how place-based AI policy can work. Pilot zones were used to test applications, build local ecosystems and create models that could later be adapted elsewhere. This approach allows policymakers to experiment before scaling up. It also gives local governments incentives to link AI to regional development goals.

However, the study finds that the equalizing effect was not uniform. It was stronger in eastern regions, in cities with more developed digital infrastructure and in areas with weaker initial human capital endowments. These differences matter because they show that AI policy outcomes depend on local conditions.

Eastern regions benefited more likely because they had stronger industrial foundations, better infrastructure, deeper supply chains and more capacity to link urban innovation with rural spillovers. Cities with stronger digital infrastructure also saw larger effects because AI requires reliable data transmission, connectivity and platform support. Without these foundations, rural AI applications may be too costly or difficult to maintain.

The human capital finding is especially notable. The policy had a stronger equalizing effect in areas with weaker human capital. One possible explanation is that AI tools can automate low-value rural tasks and allow workers to move into more productive activities. In areas where skill levels are already high, the additional equalizing effect may be smaller because urban workers also benefit strongly from AI.

For other countries, this means that AI policies should not be copied mechanically. Local context matters. Regions with weak broadband networks may need digital infrastructure before AI tools can deliver rural gains. Areas with low education levels may need training and accessible interfaces. Farming regions may need AI applications designed around local crops, terrain, climate and market structures.

The study's findings also speak directly to the Sustainable Development Goals, especially the goal of reducing inequality. AI can contribute to sustainable development only if the benefits are widely shared. That requires public policy, not just private investment. Rural households are unlikely to gain from AI if they lack connectivity, skills, financing and access to technology platforms.

The research also points to risks. AI adoption can still produce unequal outcomes if it displaces rural labor without creating new income opportunities, if large firms capture most of the data and profits, or if rural areas become dependent on technologies they cannot control. That is why policy design must include safeguards, public investment and mechanisms to ensure that rural producers retain value.

The study acknowledges its limits. It uses the urban-rural income ratio as a central measure, which captures an important part of inequality but not the full picture. Income gaps also interact with education, health, wealth, access to services and social mobility. The mechanisms examined, agricultural innovation and government attention, are important but not exhaustive. Labor mobility, access to digital services and rural entrepreneurship may also play roles.

Even with these limits, the research offers a strong message for governments in developing economies. AI should not be treated only as a tool for advanced manufacturing, surveillance, finance or urban efficiency. It can also be a rural development tool, but only if it is built into a broader strategy of inclusive growth.

• Policymakers should prioritize building digital infrastructure in rural regions. Without broadband, data platforms and reliable connectivity, AI cannot reach the communities most at risk of exclusion.
• Agricultural technology innovation must be supported through funding, pilot programs, public-private partnerships and extension services.
• Invest in rural digital skills, ensuring that farmers and rural workers can use AI tools rather than being displaced by them.
• Institutional coordination: AI policy must connect ministries and local departments responsible for agriculture, science, rural development, education, finance and digital infrastructure. Fragmented policies will limit impact.
• Local adaptation: AI solutions must match local economic structures rather than follow a one-size-fits-all model.