NTU team’s AI tool among 14 projects worldwide funded by OpenAI

The NTU team’s AI tool for economic forecasting, backed by OpenAI funding, represents a significant step in applying AI to public policy in Asia. By leveraging real transaction data from a South Korean mobile budgeting app, the system aims to provide policymakers with more accurate predictions on how economic incentives, like cash vouchers, will influence consumer behavior. This approach moves beyond traditional simulation models that often struggle to capture the complex interplay of factors like income and age, which affect human economic decisions. The real test for this unnamed tool will come in 2025, when its predictions will be validated against actual spending data from South Korea’s cash handout programs. While the US$100,000 funding, split between cash and OpenAI credits, will accelerate development, the challenge lies in ensuring the system avoids hallucination and can be generalized to diverse populations beyond its initial South Korean dataset. The goal is to have a universally applicable tool by February 2027, which could offer a new paradigm for economic planning across Asia. For Asian governments, this type of AI could refine fiscal policies, making interventions more targeted and effective. The anonymization methods used, requiring a minimum number of users to share exact characteristics, are crucial for data privacy. However, the reliance on a single, unnamed South Korean app’s data for initial training means its regional applicability will need rigorous testing and adaptation to different market dynamics and consumer behaviors across Asia.