Generative AI for Social Impact

TL;DR

The paper addresses the challenge of scaling AI for social impact by identifying three deployment bottlenecks observational scarcity data synthesis complexity and human AI alignment. It proposes a unified Generative AI approach that combines LLM agents to translate tacit expert guidance into executable objectives and diffusion models to generate realistic synthetic data and model uncertain environments, enabling robust cross domain policy transfer. Through domain illustrations on HIV prevention networks and wildlife patrol planning the authors demonstrate how generative techniques can amplify data, stabilize policy learning under nonstationarity, and align deployments with real world constraints. Embedding these generative tools into the AI4SI lifecycle promises scalable, adaptable, and trustworthy resource optimization with tangible societal benefits.

Abstract

AI for Social Impact (AI4SI) has achieved compelling results in public health, conservation, and security, yet scaling these successes remains difficult due to a persistent deployment bottleneck. We characterize this bottleneck through three coupled gaps: observational scarcity resulting from limited or unreliable data; policy synthesis challenges involving combinatorial decisions and nonstationarity; and the friction of human-AI alignment when incorporating tacit expert knowledge and dynamic constraints. We argue that Generative AI offers a unified pathway to bridge these gaps. LLM agents assist in human-AI alignment by translating natural-language guidance into executable objectives and constraints for downstream planners, while diffusion models generate realistic synthetic data and support uncertainty-aware modeling to improve policy robustness and transfer across deployments. Together, these tools enable scalable, adaptable, and human-aligned AI systems for resource optimization in high-stakes settings.