Back to projects
AI for Metascience
Project Overview
Science drives solutions to humanity’s most complex challenges. Building a meta understanding of scientific systems—how they operate, evolve, and self-regulate—is essential, particularly as AI’s role in research practices grows. In 2024, we launched the AI for Metascience project, organized around three key research threads:
- AI Adoption in Scientific Research: We monitor and analyze the integration of AI-based methods in research, aiming to construct a comprehensive view of their status, evolution, key drivers, pathways, and future directions by mining scholarly big data.
- LLM for Metascience Data Enrichment: We leverage large language models to enrich Metascience datasets, unlocking new data and perspectives for Metascience in fields such as medical research and within national funding systems, to provide more nuanced evidence and levers for science policymaking.
- LLM for Research Evaluation: We examine the validity and sensitivity of integrating LLMs into research evaluation practices, such as the peer review process. Our goal is to present a better understanding of the potential and risks associated with using LLMs in these practices.
