2023-10-18-世界银行-Novissi_Togo_利用人工智能提供应对冲击的社会保护（英）-2023.9-38页_38页_6mb

The NOVISSI program in Togo demonstrates how artificial intelligence, machine learning, and big data can revolutionize shock-responsive social protection during crises like COVID-19. It achieved rapid, digital delivery of unconditional cash transfers to over 920,000 vulnerable individuals, including informal workers and rural poor, by leveraging voter lists, geospatial data, and mobile phone metadata. Key innovations include supervised machine learning algorithms for poverty mapping at high resolution, reducing administrative burdens and enabling faster disbursements compared to traditional methods. This improved financial inclusion by accelerating mobile money adoption and minimizing leakages. However, challenges include risks of exclusion due to limited universal identification systems, a digital divide, and potential algorithmic bias not fully mitigated. Lessons emphasize the need for foundational unique IDs and dynamic social registries for inclusive, adaptive systems. Recommendations include combining digital with traditional approaches for hybrid models, enhancing data privacy through trusts, and investing in digital infrastructure to ensure equitable social protection.