Assuring AI methods for economic policymaking

Abstract

AI methods are becoming more common in the field of economics, but these models must be bias-free, fair, and explainable. In other words, we need AI assurance. Economic forecasting has benefited from machine learning techniques, such as neural networks, to increase model performance, but these AI techniques must be audited, accountable, and interpretable to be useful for economic policymaking. The rise of natural language processing and large language models has created new challenges for economic policymaking institutions, which need to be aware of AI assurance and how to harness them safely. https://www.sciencedirect.com/science/article/abs/pii/B9780323919197000251?via%3Dihub