In recent years, technological innovation has exhibited unprecedented complexity and uncertainty, and traditional frontier technology monitoring systems have repeatedly failed to anticipate disruptive transformations. The explosive emergence of ChatGPT, for example, revealed its revolutionary impact, yet it was scarcely predicted by mainstream analytical frameworks prior to its debut. This highlights the limitations of innovation monitoring methods based on explicit indicators such as academic papers and patents when confronted with increasingly latent and fragmented emerging technologies. As the geometry of innovation continues to undergo profound restructuring, building a monitoring system adapted to the new innovation ecosystem has become a critical issue for safeguarding national technological competitiveness and industrial security. A high-sensitivity, early-warning intelligence monitoring framework can not only identify potential directions of frontier innovation but also provide timely innovation intelligence for government and corporate decision-makers, enabling them to make precise frontier innovation deployments and resource allocations in the face of fierce global technological competition.

The explosion of large language models technology represented by ChatGPT has refreshed our understanding of artificial intelligence technology, impacted and reconstructed many traditional industries. This article focuses on the impact of large language models on intelligence analysis work, compares the two processes of intelligence analyst training and large language model construction, analyzes the similarities and differences between intelligence analysts and large language model. The study finds that large models are more like anthropomorphic intelligence analysis tools, but intelligence analysts have characteristic advantages that cannot be replaced by current large language models, such as emotions, sensations, bounded rationality, flexibility, and discernment. At the same time, the application of large language models in intelligence work will lead to the restructuring of the intelligence analysis process. The restructured process will still be dominated by intelligence analysts, but large language models will be involved in each stage to varying degrees.

This paper enriches the theoretical foundation for identifying enterprise competitors, expands the methods for identifying enterprise competitors, and provides reference for enterprises to analyze the competitive situation. Firstly, the VRIO model is introduced to design a “target enterprise competitor” technology competition gap measurement index based on patents from four aspects: value, scarcity, inimitability and organization. Secondly, classifies competitors into single type, cross type, combined type, and comprehensive type, and further construct a comprehensive competition index indicator to achieve inter group comparison of competitors of the same type of competitors under the same dimension. Finally, refine the granularity of competitor analysis and analyzes the technological competitive elements of competitors in detail. Targeting FY company in the drone field, identifies its competitors. The results indicate that the enterprise competitor identification method guided by the VRIO model enriches the theoretical support for enterprise competitor identification. It also assists target enterprises objectively recognize the multidimensional resource gap between themselves and their competitors, provides decision-making support for leveraging their strengths and avoid weaknesses, and formulate technology layout strategies.

In today’s globalized economic environment, the trend of rapid development in fintech is evident. The widespread application of technologies such as blockchain, artificial intelligence, and big data in the financial sector poses unprecedented challenges to financial security. With the accelerated iteration of artificial intelligence technology and its explosive growth in applications, the consequences of tail risks, if they occur, will be catastrophic. As an effective analytical tool, the intelligence methods help identify and predict potential tail risks in the process of artificial intelligence application. By early detection of these risks, we can take timely measures to respond, thereby ensuring the robustness and security of the financial system and safeguarding the healthy development of China’s financial industry.

Quantitative evaluation of science and technology (S&T) policies implementation helps assess effectiveness of current policies and provides a data basis for future policymaking. This study examines 81 policies documents related to the “50 New S&T Policies” issued in Zhejiang Province from 2019 to 2022. Using content analysis, it evaluates implementation quantitatively from the dimensions of policy level, policy tools, and policy content. Findings show significant differences in the number of policies and public announcements between provinces and cities in terms of policy level. There are also significant differences, in terms of policy tools, in the average distribution of supply-oriented tools, demand-oriented tools, and environmental-oriented tools; In terms of policy content, the supply-oriented policies, in the major policy categories, account for the highest proportion, reaching 48%, and the policy subcategories of public services and capital investment have a relatively high proportion. Finally, the study suggests expanding the coverage of policies formulation and introduction, enhancing the continuous implementation level of policies, strengthening the content of demand-oriented policies, and improving the integrated form and measures of policies.

How to coordinate the prospective governance of serving both development and security of science and technology has become an urgent issue to be solved for the high-quality development of scientific and technological decision-making consultation services. Taking the Center for Security and Emerging Technology (CSET) at Georgetown University as an example, this paper explores the operational mechanism of integrating and utilizing open-source intelligence to serve the prospective governance of development and security of science and technology. It also puts forward suggestions from aspects of strengthening the top-level design of strategic intelligence services, enhancing the ability to discover scientific and technological information resources, promoting collaborative innovation with internal and external institutions, and optimizing the management system for the transformation of research achievements, providing references for the institutions of scientific and technological intelligence in China transforming into think tanks and serving development and security of science and technology more effectively.

As an important content of brain science, brain atlas is the basis for studying the structure, function and regulation of the brain. China and the United States have launched national brain science programs, which brain atlas research has been taken as the focus. In addition, this paper uses bibliometrics and content analysis to benchmark the basic research in China and the United States. The results show that under the orderly guidance of the strategic layout, the brain atlas research in the United States has successively completed cell classification and typing, brain atlas of different species, etc., while the brain atlas research in China lacks of a step-by-step and goal-by-goal strategic layout. In terms of basic research, there is a significant gap between China and the United States, with the United States having strong scientific research strength and world-class research results in brain atlas research. In China, the research on brain atlas is basically in the early stage of development, and the research force has not yet been formed, the research results need to be accumulated urgently. Through benchmarking, two implications are formed for the future rapid consolidation of brain atlas research foundation in China.