From Automation to Autonomy: A Survey on Large Language Models in Scientific Discovery

2025/05/19 by Tianshi Zheng, Zheng, Tianshi, Zheye Deng +11 · 1 voice · 21 citations
Computer Science · #Computation and Language (cs.CL) #FOS: Computer and information sciences #cs.CL

paper · pdf · doi:10.48550/arxiv.2505.13259

arxiv published 2025/05/19 · arxiv updated 2025/09/17

Abstract

Large Language Models (LLMs) are catalyzing a paradigm shift in scientific discovery, evolving from task-specific automation tools into increasingly autonomous agents and fundamentally redefining research processes and human-AI collaboration. This survey systematically charts this burgeoning field, placing a central focus on the changing roles and escalating capabilities of LLMs in science. Through the lens of the scientific method, we introduce a foundational three-level taxonomy-Tool, Analyst, and Scientist-to delineate their escalating autonomy and evolving responsibilities within the research lifecycle. We further identify pivotal challenges and future research trajectories such as robotic automation, self-improvement, and ethical governance. Overall, this survey provides a conceptual architecture and strategic foresight to navigate and shape the future of AI-driven scientific discovery, fostering both rapid innovation and responsible advancement. Github Repository: https://github.com/HKUST-KnowComp/Awesome-LLM-Scientific-Discovery.

Cited by

• VASP Agent: An Agentic Framework for Autonomous First-principles Calculations
• Synthelite: Chemist-aligned and feasibility-aware synthesis planning with LLMs
• Beyond Text-to-SQL: Autonomous Research-Driven Database Exploration with DAR
• DataGovBench: Benchmarking LLM Agents for Real-World Data Governance Workflows
• Knowledge-Informed Automatic Feature Extraction via Collaborative Large Language Model Agents
• ATLAS: A High-Difficulty, Multidisciplinary Benchmark for Frontier Scientific Reasoning
• Deep Ideation: Designing LLM Agents to Generate Novel Research Ideas on Scientific Concept Network
• ToxScreen: Detecting Whether an LLM Has Been Poisoned
• CritiCal: Can Critique Help LLM Uncertainty or Confidence Calibration?
• AutoGraph-R1: End-to-End Reinforcement Learning for Knowledge Graph Construction
• AlphaQuanter: An End-to-End Tool-Orchestrated Agentic Reinforcement Learning Framework for Stock Trading
• GenCellAgent: Generalizable, Training-Free Cellular Image Segmentation via Large Language Model Agents
• DixitWorld: Evaluating Multimodal Abductive Reasoning in Vision-Language Models with Multi-Agent Dixit Gameplay
• Autonomous Agents for Scientific Discovery: Orchestrating Scientists, Language, Code, and Physics
• AutoPR: Let's Automate Your Academic Promotion!
• SciSchema.org: A Multidisciplinary Collection of Schemas for Structured Scientific Process Descriptions
• CLVisc Agent for autonomous relativistic hydrodynamics studies
• RELISH: LLM REgression with a Latent Iterative State Head
• Structuring the Unstructured: A Systematic Review of Text-to-Structure Generation for Agentic AI with a Universal Evaluation Framework
• Rethinking Prospect Theory for LLMs: Revealing the Instability of Decision-Making under Epistemic Uncertainty
• How Far Are AI Scientists from Changing the World?

Discussions

• From Automation to Autonomy [hn, 2 points, 1 comments]

Related

• Evolving Roles of LLMs in Scientific Innovation: Assistant, Collaborator, Scientist, and Evaluator
• Autonomous Agents for Scientific Discovery: Orchestrating Scientists, Language, Code, and Physics
• Advancing the Scientific Method with Large Language Models: From Hypothesis to Discovery
• Transforming Science with Large Language Models: A Survey on AI-assisted Scientific Discovery, Experimentation, Content Generation, and Evaluation
• AI-Researcher: Autonomous Scientific Innovation