Iqbal Dhaliwal
Global Executive Director
The Abdul Latif Jameel Poverty Action Lab (J-PAL)
The key is in how organizations leverage AI. All of us see the transformative potential of AI, but fewer consider its boundaries. While AI can be powerful, it can’t replace the unique skills, judgment, and human connection that people bring to their work––it’s a tool for those people to be most effective.
Take AI-driven advice for farmers in rural Ghana or India for example. Are we sure that it’s genuinely better than the guidance they would get from local agricultural extension workers? Even if the advice is good, are farmers willing to trust AI over a real person? It’s critical that we measure these outcomes and adjust accordingly—especially for the communities whose lives we aim to improve.
Similarly, AI has the potential to make many people’s work more effective, efficient, and rewarding, but we need to ensure economic and tax policy does not create artificial incentives for AI that pushes them aside.
As AI tools become increasingly commonplace, we must rigorously evaluate AI. Causal evidence on AI’s societal impacts is currently limited, leaving companies, social innovators, and policymakers without information to maximize the benefits of AI while mitigating its risks. We have decades of experience on how randomized control trials can be used to measure impact of social programs, including tech-infused innovations. We have a responsibility to rigorously study how this new technology can help or harm people across multiple dimensions including income, but also metrics of well being like health and empowerment.
These insights will equip decision-makers with the information they need to design and implement AI solutions that genuinely benefit people, especially those from underserved communities.