From agri to road safety to public health, AI-driven governance can make lives easier: Michael Kremer
Kremer, who was awarded the 2019 Nobel Prize in Economic Sciences, jointly with Abhijit Banerjee and Esther Duflo, for their experimental approach to alleviating global poverty, had been invited to share his insights on AI integration into governance.
Artificial intelligence (AI), if embedded within evidence-based policymaking, can transform agriculture, road safety, education and public health, but only if governments actively shape its deployment, Nobel laureate economist Michael Kremer said during a public lecture at Mahatma Gandhi State Institute of Public Administration on Friday.
Kremer, who was awarded the 2019 Nobel Prize in Economic Sciences, jointly with Abhijit Banerjee and Esther Duflo, for their experimental approach to alleviating global poverty, had been invited to share his insights on AI integration into governance.
The event, titled ‘Economics and policy innovation in India’, was organised by the Centre for Research in Rural and Industrial Development (CRRID) and the Chandigarh Citizens Foundation, a collaborative platform and a local think tank for tricity.
AI-driven forecast for farmers
Speaking about how AI can make life better for people, Kremer cited an ongoing research on how it is helping farmers in India. “Timely and comprehensible AI driven weather forecasts can significantly influence farm decisions. In one north Indian state, farmers who received advance forecasts were 10% more likely to raise nurseries and 12% more likely to adopt hybrid seeds,” he said, demonstrating how improved information can shape risk-taking and productivity. He stressed that forecasts should go beyond rain predictions to include “soil moisture, humidity, night heat and wind speed, factors critical for crops and livestock.”
Automated enforcement to make roads safer
{{/usCountry}}Automated enforcement to make roads safer
{{/usCountry}}Highlighting India’s annual toll of over 1.5 lakh road crash deaths, Kremer pointed to AI-enabled driver testing and enforcement systems as tools to improve safety and transparency. He cited automated driving test monitoring platforms that objectively evaluate driver performance. “Comparative analysis suggested a 20-30% reduction in reported unsafe driving, particularly among young drivers, in areas where the system was introduced,” he said. Such systems, he noted, can “reduce discretion and improve public trust in government processes.”
{{/usCountry}}Highlighting India’s annual toll of over 1.5 lakh road crash deaths, Kremer pointed to AI-enabled driver testing and enforcement systems as tools to improve safety and transparency. He cited automated driving test monitoring platforms that objectively evaluate driver performance. “Comparative analysis suggested a 20-30% reduction in reported unsafe driving, particularly among young drivers, in areas where the system was introduced,” he said. Such systems, he noted, can “reduce discretion and improve public trust in government processes.”
{{/usCountry}}AI-based highway surveillance can also detect speeding, wrong-side driving and seatbelt violations. Kremer said AI makes it possible to shift enforcement models from “high fines with low detection probability” to “modest penalties with high-detection rates,” which could be both fairer and more effective.
{{/usCountry}}AI-based highway surveillance can also detect speeding, wrong-side driving and seatbelt violations. Kremer said AI makes it possible to shift enforcement models from “high fines with low detection probability” to “modest penalties with high-detection rates,” which could be both fairer and more effective.
{{/usCountry}}Water interventions could avert over 1 lakh child deaths
Presenting findings from a meta analysis of 19 randomised evaluations, Kremer said water-treatment interventions can reduce under-five mortality by 21%. Universal rural access to treated water in India could “avert over one lakh child deaths annually,” he said, describing such interventions as “highly cost-effective.” He suggested that AI-enabled sensors and performance-linked procurement could help reduce costs of water-treatment systems.
Improving learning outcomes
Kremer said education technology has demonstrated very large gains in learning outcomes, but only when actively used and monitored. In one government programme, software usage was initially negligible. When authorities began tracking login data and contacting schools with low utilisation, usage increased to about an hour per week, resulting in “substantial improvements in learning outcomes, especially among younger students and girls.”
“The challenge is not just technology, but integrating it into government systems,” he said, proposing automated monitoring dashboards to flag low usage to administrators.
Is it feasible in India?
During the interactive session, Amrita Shergill, former chairperson of the department of economics, Panjab University, questioned the feasibility of AI adoption in an economy like India, where only about 30% of employment lies in services. Responding to this, Kremer said India’s digital public infrastructure and expanding AI ecosystem position it to align technological innovation with inclusive development. He added that embedding evaluations within government rollouts allows rigorous impact assessment without increasing costs. Acknowledging anxieties around jobs, he said technological transitions, if guided by policy, can strengthen public service delivery.
