AI AGENT IN SCI-TECH INTELLIGENCE ANALYSIS: CURRENT APPLICATIONS AND FUTURE TRENDS
Objective: In the data-intelligent era, sci-tech intelligence analysis faces the dual challenges of information overload and insufficient depth in knowledge mining. Driven by large language models, AI Agents—as new autonomous actors capable of perception, planning, execution, and memory—are profoundly reshaping the intelligence workflow paradigm. This paper aims to systematically analyze the current applications, core capabilities, and development trends of AI Agents in sci-tech intelligence analysis, providing a reference for theoretical evolution and innovative practice in information science. Process: Using literature review and logical deduction, this paper first constructs an analytical framework encompassing four elements: "human–agent–information–technology." It then examines the core capabilities of AI Agents, from planning and memory to multi-agent collaboration. On this basis, it delves into their current applications in intelligence perception, organization, and generation. Finally, it discusses emerging trends such as human–AI collaboration paradigms and knowledge production transformation, along with potential challenges. Conclusion: AI Agents are driving the transformation of sci-tech intelligence analysis from "human-in-the-loop" to "human-on-the-loop," enabling intelligent and pipeline-based intelligence production processes. In the future, intelligence work will move toward a human–multi-agent collaborative paradigm, yet faces risks such as AI hallucinations and cognitive offloading, requiring urgent countermeasures at technical, literacy, and institutional levels.