Stable Menus of Public Goods: AI-Enabled Progress

The study investigates AI workflows in the context of economic research, specifically analyzing the "Stable Menus of Public Goods" problem. It evaluates the impact of human intuition in prompts and multi-turn interactions on LLM performance, suggesting that these elements enhance model effectiveness. However, the findings indicate that the LLM slightly underperforms compared to a first-year PhD student, highlighting the nuanced role of human expertise in complex economic tasks.