AI Economist Agent: An Agentic Framework for Model-Grounded Economic Analysis with RAG, Knowledge Graphs, and Large Language Models
Researchers have developed an AI economist agent designed to ground economic analysis using large language models, knowledge graphs, and retrieval-augmented generation. This framework aims to ensure economic claims are supported by theory and data, rather than relying solely on LLM outputs. The AI economist agent has been evaluated in generating reports on U.S. inflation persistence and Federal Reserve policy, as well as in creating narratives for U.S. commercial real estate refinancing stress tests, demonstrating improved economic coherence and traceability. AI
IMPACT This framework could enhance the reliability and traceability of AI-generated economic analyses, making them more useful for policy and financial stress testing.