Amid ongoing debates regarding the regulation of artificial intelligence, the World Bank has argued that the technology holds significant potential to advance global development and reduce inequality. While concerns about AI exacerbating existing disparities are widespread, officials at the institution believe the tools can be harnessed for positive social impact.
Paschal Donohoe, Managing Director and Chief Knowledge Officer at the World Bank Group, highlighted concrete benefits observed in key sectors such as agriculture, education, and healthcare. He noted that AI can substantially improve crop productivity by assisting farmers in managing planting and harvesting processes, thereby increasing their incomes.
Donohoe cited a specific example from India, where small-scale rice farmers utilized AI to receive tailored advice on cultivation and harvest techniques. “I saw a very clear example of this in India, where I saw small rice farmers who, through use of artificial intelligence, were given better advice on how rice could be planted and harvested,” he said.
The World Bank official emphasized that developing economies can benefit from AI if the technology is adapted to local conditions—a concept he described as “small AI.” This approach involves systems that are adjusted to specific local circumstances, capable of operating offline, and functional in environments with limited connectivity or electricity.
Donohoe stressed that development agencies must remain focused on fundamental needs, including access to clean water, healthcare, and education, even as daily life becomes increasingly digital. He argued that while AI presents challenges, it can play a positive role when guided by appropriate policy frameworks.
“We are recognising the change that is now taking place with AI, and we are making the case that while of course it brings challenges, small AI in particular, harnessed correctly with the right policies around it, can play a positive role,” Donohoe stated.
I worry about who builds these tools. Will they reflect local realities, or just replicate Western biases in a cheaper package? Implementation details matter.
The India rice farmer example is compelling. If offline AI can actually boost yields without internet, that could be a real game-changer for rural communities.