The Rise of Large Language Models and the Direction and Impact of US Federal Research Funding

• https://arxiv.org/abs/2601.15485
• Yifan Qian, Zhe Wen, Alexander C, Yue Bai, Erzhuo Shao, Dashun Wang

LLMs, Science policy, Research funding, Science of science

From Dashun Wang‘s group as Furnas2026geopolitics and Shao2025sciscigpt, this paper examines how the rise of LLMs has reshaped federal research funding priorities and their downstream impact on science.

The key idea is to use AI to rewrite pre-chatGPT abstracts to create a positive sample and build a classifier for assigning an AI score to grant proposal. AI-assisted proposals are more likely to get funding but more similar to previously funded projects (less novel?)