Agentic AI for Scientific Discovery: A Survey of Progress, Challenges, and Future Directions

2025-03-12

AI Summary

This study addresses the systematic integration of Agentic AI into scientific discovery, targeting automation and collaborative intelligence across literature review, hypothesis generation, experimental design, and result analysis. Methodologically, it introduces the first taxonomy of Agentic AI for scientific discovery; establishes a human-AI co-calibration mechanism, reliability assessment framework, and ethical governance principles to formalize the “automation → augmented intelligence” progression; and integrates large language models, planning-and-reasoning engines, tool-use interfaces, scientific knowledge graphs, and domain-specific APIs—validated on multi-stage benchmarks and real-world research workflows. Contributions include: (1) a synthesis of over 100 representative works; (2) unified evaluation metrics and an open-source dataset inventory; (3) identification of five core challenges; and (4) a pragmatic, interdisciplinary research roadmap for advancing Agentic AI in science.

Technology Category

• Humans and AI: Planning and Decision Support for Human-Machine Teams
• Multiagent Systems: Agent/AI Theories and Architectures
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Research challenges in human and human-AI computation

Abstract

The integration of Agentic AI into scientific discovery marks a new frontier in research automation. These AI systems, capable of reasoning, planning, and autonomous decision-making, are transforming how scientists perform literature review, generate hypotheses, conduct experiments, and analyze results. This survey provides a comprehensive overview of Agentic AI for scientific discovery, categorizing existing systems and tools, and highlighting recent progress across fields such as chemistry, biology, and materials science. We discuss key evaluation metrics, implementation frameworks, and commonly used datasets to offer a detailed understanding of the current state of the field. Finally, we address critical challenges, such as literature review automation, system reliability, and ethical concerns, while outlining future research directions that emphasize human-AI collaboration and enhanced system calibration.

Problem

• Explores Agentic AI's role in automating scientific discovery processes.
• Surveys progress, tools, and challenges in AI-driven scientific research.
• Addresses reliability, ethics, and future human-AI collaboration in science.

Innovation

• Agentic AI automates scientific research processes.
• AI systems enable autonomous hypothesis generation and testing.
• Survey categorizes tools and datasets for scientific discovery.