Is Innovation Becoming More Conservative? An LLM Framework for Quantifying Risk

DOI

10.2991/978-94-6239-787-3_35How to use a DOI?

Keywords

Innovation; Large language Models; Risk; Novelty

Abstract

Despite growing research investment, the pace of scientific innovation has slowed markedly, raising concerns among policymakers and scholars. Existing bibliometric approaches to measuring recombination-based innovation rely on static word embeddings and computationally expensive classifiers, limiting their applicability at scale. We propose a co-citation prediction framework that replaces Word2Vec with DeCLUTR—a contrastive-learning pretrained language model—and substitutes SVM with a single-hidden-layer MLP, achieving state-of-the-art classification performance while reducing inference time by orders of magnitude. Validated against a large-scale survey of researchers, our innovation index substantially outperforms the prior baseline in correlating with self-reported overall research risk in the medical domain. Applied to millions of medical papers from Web of Science across eight major countries over nearly two decades, we find a consistent decline in recombination-based innovation, supporting the hypothesis that scientific research is becoming more conservative. Top-tier institutions consistently outperform national averages, and among Chinese cities, established research centers exhibit stable trends while emerging innovation hubs show greater volatility.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Zhao Wu
AU  - Fuchuan Qu
AU  - Deyun Yin
AU  - Sotaro Shibayama
PY  - 2026
DA  - 2026/09/29
TI  - Is Innovation Becoming More Conservative? An LLM Framework for Quantifying Risk
BT  - Proceedings of the 2026 4th International Conference on Management Innovation and Economy Development (MIED 2026)
PB  - Atlantis Press
SP  - 343
EP  - 351
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6239-787-3_35
DO  - 10.2991/978-94-6239-787-3_35
ID  - Wu2026
ER  -