主管:中国科学院
主办:中国科学院文献情报中心、中国高科技产业化研究会
ISSN:1006-222X
CN:11-3556/N
Home Table of Contents

09 March 2026, Volume 32 Issue 2
    

  • Select all
    |
  • LIU Cancan, XIANG Xiaowen
    High-Technology and Commercialization. 2026, 32(2): 7.
    Abstract ( )   Knowledge map   Save
  • QIU Rui, XIANG Xiaowen
    High-Technology and Commercialization. 2026, 32(2): 9.
    Abstract ( )   Knowledge map   Save
  • ZHU Zhaocheng, XU Binxiang, LUO Hongwei
    High-Technology and Commercialization. 2026, 32(2): 12.
    Abstract ( )   Knowledge map   Save
    Spatial intelligence represents the cutting-edge of artificial intelligence development. It is not only the product of the evolution of multiple digital technology paradigms but also an algorithmic tool that drives multiple changes in social governance. At present, public security organs have not yet fully integrated spatial intelligence into their practical work. Grassroots police officers mostly adopt traditional models to handle police situations, which are cumbersome and inefficient.Studies have shown that spatial intelligence can play a good auxiliary role for public security organs in reproducing crime scenes, constructing evidence chains, and daily patrols. However, problems such as privacy leakage, technical black boxes, and over-reliance may arise during its use. Therefore, methods such as strengthening privacy protection, optimizing technology, and formulating corresponding laws have been proposed to solve these problems.Spatial intelligence empowers traditional policing to move towards a new era of smart policing characterized by digitalization, networking, and intelligence. It promotes the intelligent transformation of policing activities from being “people-oriented” to “technology-led”, and accelerates the modernization of policing activities.
  • YIN Qiang
    High-Technology and Commercialization. 2026, 32(2): 16.
    Abstract ( )   Knowledge map   Save
    Domestic large-scale models represented by DeepSeek, with their powerful capabilities in semantic understanding, knowledge generation, and logical reasoning, are reshaping the innovative pathways and methodological systems of academic research. To systematically reveal the research hotspots and development trends in this field, this study conducted a quantitative analysis of 236 core DeepSeek-related articles from CNKI)using the LDA topic model. By refining the topic terms, four key research themes were identified: intelligent information management and cultural service transformation; education and social change in the intelligent era; the digital economy and technology industry ecosystem; and technological risks and global governance. Based on the refined themes, a visual analysis was performed, followed by future prospects.
  • ZHANG Lingli
    High-Technology and Commercialization. 2026, 32(2): 22.
    Abstract ( )   Knowledge map   Save
    Cosmic dust is a key material connecting stellar evolution, planetary formation, and the evolution of the interstellar medium. Benefiting from the motion of the solar system relative to the local interstellar cloud, several deep-space exploration missions have achieved in-situ detection of interstellar dust. Impact ionization detectors such as those on Helios, Ulysses, Galileo, and Cassini, along with the Stardust sample return mission, have made significant progress in analyzing the dynamical characteristics and composition of particles, initially revealing the mass distribution, radiation pressure response, and material composition of interstellar dust. However, due to limitations in orbital configuration, detection geometry, instrument performance, and dynamic filtering effects, significant uncertainties remain in existing observations regarding the particle size range, compositional completeness, and understanding of internal structure, requiring further breakthroughs.
  • WANG Linyu
    High-Technology and Commercialization. 2026, 32(2): 26.
    Abstract ( )   Knowledge map   Save
    Fishing ban policies are vital for biodiversity conservation, yet traditional manual patrols are limited by low efficiency and coverage. With advances in artificial intelligence and video surveillance, a collaborative “human defense + technological defense” model has emerged. This study examines its theoretical basis, key mechanisms, and promotion pathways. The findings show that the model improves supervisory precision and coordination, supporting biodiversity protection and ecological governance modernization.
  • CAI Yixuan
    High-Technology and Commercialization. 2026, 32(2): 29.
    Abstract ( )   Knowledge map   Save
    Against the backdrop of accelerated digital and intelligent transformation, small and medium-sized retail enterprises are confronted with prominent contradictions in human resource management due to the industry characteristics of “scattered stores, frequent personnel turnover, and high labor costs”. This paper adopts a hybrid research method of “questionnaire + interview + multi-case analysis”, based on 296 valid questionnaires, in depth interviews with 23 stores and 12-month tracking data of 3 typical enterprises, systematically analyzes the current application status of digital tools and the bottleneck of efficiency transformation, and constructs a three in-one optimization strategy system of “precise empowerment by modules + full-process collaboration + efficiency transformation”. Empirical verification shows that this system can effectively solve core problems such as data silos and help enterprises improve the quality and efficiency of human resource management.
  • YANG Jianyun
    High-Technology and Commercialization. 2026, 32(2): 32.
    Abstract ( )   Knowledge map   Save
    In the digital age, libraries, archives, and museums house vast amounts of high-quality information resources, yet they face challenges such as heterogeneous fragmentation, weak connections, and inefficient utilization. As a core technology of semantic networks, knowledge graphs possess strong semantic expression and association-mining capabilities. This paper addresses the practical difficulties in integrating information resources from libraries, archives, and museums, explains the characteristics of knowledge graph technology, analyzes its application value, and proposes a comprehensive integration strategy that includes ontology construction, data fusion, platform establishment, and collaborative management. This strategy aims to provide a technical path and practical reference for achieving semantic interconnection of resources and intelligent upgrading of services.
  • LIN Hang
    High-Technology and Commercialization. 2026, 32(2): 35.
    Abstract ( )   Knowledge map   Save
    As the underlying core technology of the new round of technological revolution, Artificial Intelligence Generated Content (AIGC) has distinct characteristics such as automation and intelligence, personalization and customization, diversity and innovation, as well as real-time and efficiency. The deep application of AIGC technology in the field of higher education will reshape the teaching mode of university education, construct a personalized teaching system, stimulate the innovative vitality of university education, and enhance the response quality and efficiency of university education. As a result, the future functions of universities will evolve towards “learning centers”, and university education will undergo changes towards interdisciplinary and general education curriculum systems, experiential and transformative service outputs, intelligent and cross age learning environments. Faced with the wave of higher education reform driven by AIGC technology, Chinese universities need to build interdisciplinary curriculum systems, establish intelligent teaching platforms and resource libraries, promote collaborative cooperation between industry, academia and research, and strive to enhance teachers’ technological literacy.
  • MA Zuyi
    High-Technology and Commercialization. 2026, 32(2): 42.
    Abstract ( )   Knowledge map   Save
    The 20th CPC Central Committee's fourth plenum called for seizing tech revolution opportunities to develop new productive forces. While AI and robots boost efficiency, they also disrupt employment, erode tax bases, and widen income gaps, necessitating robot taxation for fairness. Key obstacles include unclear legal status of robots, ambiguous taxpayer identity and tax nature, and lack of collection systems. This paper suggests optimizing the current tax system by clarifying taxpayer status and tax nature, and improving collection mechanisms to integrate robot-related activities, balancing tech progress with public interest.
  • CUI Guannan, WANG Ying
    High-Technology and Commercialization. 2026, 32(2): 45.
    Abstract ( )   Knowledge map   Save
    When digital technology is deeply integrated with public governance, artificial intelligence has become a key driving force for the transformation of public hospital governance and the upgrade of intelligent services. This article takes public value creation and sustainability as the two main lines, focusing on the practical logic of AI empowering intelligent services in public hospitals, analyzing its paths of public value creation in enhancing the accessibility of medical services, optimizing the allocation of public medical resources, improving service quality and patient experience, and dissecting the core institutional obstacles such as budget investment, personnel adaptation, and data barriers faced by AI empowerment. It proposes corresponding sustainable optimization paths. The main research focus of this article is to solve the institutional predicaments related to the release of public value in intelligent services, release the public value of intelligent services, and assist public hospitals in achieving high-quality development.
  • FENG Xiaoyan
    High-Technology and Commercialization. 2026, 32(2): 48.
    Abstract ( )   Knowledge map   Save
    Using statistical analysis to mine topic information as input for generative models improves the interpretability of academic topic selection. According to manual preliminary directions, literature is retrieved from CNKI, processed by topic analysis, feature dimensionality reduction, topic combination mining and multi-dimensional recommendation, and then titles are generated by the model.Compared with direct generation, this method produces more specific and in-depth topics, and the statistical process improves interpretability. However, limitations exist: step-by-step selection is only at the process level, not integrated into the model; user preferences are ignored; and results lack expert verification. In general, step-by-step topic selection combining statistics and generative technology greatly improves reliability and interpretability.
  • LIU Pengbo, CHEN Qimei, LI Haoze, WANG Tangrong, YAO Shanyun
    High-Technology and Commercialization. 2026, 32(2): 51.
    Abstract ( )   Knowledge map   Save
    The rapid advancement of artificial intelligence technology makes it crucial to understand its global development trends for technological innovation and industrial planning. Utilizing global AI patent data, this study performs a quantitative analysis across three dimensions: macro-trends, frontier research directions, and emerging application scenarios. Findings indicate that global AI patenting entered a phase of exponential growth in 2013; while China leads in total patent volume, the United States maintains a competitive edge in high-barrier vertical domains. Currently, algorithm optimization, model lightweighting, and multimodal fusion constitute the primary technological frontiers. Furthermore, application scenarios such as the Metaverse, the low-altitude economy, and green/low-carbon initiatives have gained significant momentum recently. Research indicates that global AI innovation is transitioning from scale expansion to efficiency optimization and scenario-driven development.
  • WANG Cheng, ZHANG Lingkai, WANG Bo
    High-Technology and Commercialization. 2026, 32(2): 56.
    Abstract ( )   Knowledge map   Save
    With the ongoing deepening of global digitalization and intelligentization, the satellite navigation industry has become a critical infrastructure closely tied to national strategic security and economic core competitiveness. Currently, the global navigation satellite system (GNSS) landscape is characterized by a “four-power coexistence” of the U.S. GPS, China’s BeiDou, Russia’s GLONASS, and the European Union’s Galileo. As a latecomer, the BeiDou system has achieved global service capabilities and deep penetration in the domestic market, yet it faces multiple challenges at the international level, including ecosystem lock-in by GPS, geopolitical interference, and competition over standards. Therefore, a systematic study of BeiDou’s internationalization path is essential for enhancing its global competitiveness, safeguarding national spatiotemporal information security, and promoting the high-quality “going global” strategy of the aerospace industry. By analyzing the global competitive landscape and the strategies of major systems, this research assesses the foundation and challenges for BeiDou’s internationalization, ultimately proposing a comprehensive and differentiated advancement path that prioritizes technology and standards, fosters collaborative development of the application ecosystem, and expands market segments in a layered manner.
  • Lamucao
    High-Technology and Commercialization. 2026, 32(2): 61.
    Abstract ( )   Knowledge map   Save
    Virtual Reality (VR) technology is profoundly reshaping the development pattern of the tourism industry. By constructing immersive digital experience scenarios, VR technology breaks the temporal and spatial constraints of traditional tourism and injects new growth drivers into the tourism industry. At present, the integrated development of virtual and real tourism is confronted with core challenges such as inadequate technological adaptability, immature business models and fragmented user experience. Based on an analysis of industrial practices, this paper proposes integrated development paths including the construction of a technological standardization system, the deepening of scenario-based applications and the provision of data-driven precise services. It aims to promote the deep coupling of VR technology and the tourism industry, and realize the coordinated development of technological empowerment and industrial upgrading.
  • WANG Guangcun, ZHAO Yanbing, XU Xiujie, WANG Guofeng
    High-Technology and Commercialization. 2026, 32(2): 64.
    Abstract ( )   Knowledge map   Save
    In the face of the rigid constraints of the domestic “carbon neutrality” strategy and the continuous escalation of global green trade barriers, this paper aims to integrate existing research results and construct a systematic response framework for the green transformation of the discrete manufacturing industry, driven by three engines: “governance foundation building, technological innovation, and market empowerment”. The research shows that discrete manufacturing enterprises must abandon the fragmented technological renovation mindset and instead adopt a systematic strategic approach driven by the three engines.
  • QU Minling, LIANG Haiqiong, TIAN Senfang, LV Jiabei, LU Cuiyi
    High-Technology and Commercialization. 2026, 32(2): 67.
    Abstract ( )   Knowledge map   Save
    Against the backdrop of the “dual carbon” goals, this study focuses on the Xizang. By developing an integrated evaluation index system for the digital economy, ecological environment, and clean energy systems, we conduct an empirical analysis using panel data (2013–2023), employing the TOPSIS entropy weight method and a coupling coordination degree model. The results indicate: (1) The overall development level of the three systems demonstrates an upward trajectory, characterized by leapfrog development in the digital economy, continuous optimization of clean energy, and an inverted U-shaped curve in the ecological environment. (2) The coupling coordination degree of the three systems has undergone a three-stage evolutionary process, with synergistic effects progressively intensifying.
  • HE Bing, WANG Xiaohang, LANG Hong, JIANG Long, LI Chao
    High-Technology and Commercialization. 2026, 32(2): 71.
    Abstract ( )   Knowledge map   Save
    This paper systematically reviews the experiences and lessons from Japan’s regional revitalization practices across three distinct periods. It identifies key characteristics of Japan’s IUR collaborative innovation system at three levels: legal breakthroughs, policy innovations, and governance transformations. Drawing on open innovation theory, the study categorizes the resulting innovation platforms into three types: government-led, regionally embedded, and market-driven. Based on this analysis, implications are derived for deepening IUR collaboration within China’s regional economic revitalization efforts. These include that restructuring the institutional core by establishing a framework for service invention rights reform and a regional innovation compatibility assessment mechanism; reshaping structural resilience through legislation ensuring SME participation in R&D projects and the creation of technology adaptation centers; rebuilding ecological ethics by incorporating traditional culture preservation metrics into evaluation systems and establishing innovation insurance funds.
  • ZHU Chonggao, CHENG Ling
    High-Technology and Commercialization. 2026, 32(2): 76.
    Abstract ( )   Knowledge map   Save
    Against the backdrop of the current complex and volatile international landscape, establishing a highly symbiotic innovation ecosystem based on International Joint Research Centers is highly beneficial. This approach will further enhance the innovation capacity and sustainable development capabilities of China’s research institutions. This paper employs the TOE (Technology-Organization-Environment) framework and builds upon the concepts of symbiosis and innovation ecology. It systematically examines the factors influencing the strength of symbiosis within innovation ecosystems from technological, organizational, environmental, and safeguard mechanism perspectives. It is demonstrated that constructing a highly symbiotic innovation ecosystem using the TOE framework can significantly improve the R&D capabilities, management efficacy, and resource integration capacity of International Joint Research Centers. This paper also explores approaches for building such symbiotic innovation ecosystems within International Joint Research Centers, including specific implementation strategies and relevant policy recommendations, providing references for enhancing international science and technology management practices.
  • PU Xue
    High-Technology and Commercialization. 2026, 32(2): 81.
    Abstract ( )   Knowledge map   Save
    This study examines 43 policy documents on science and technology achievement transformation issued by Tianjin Municipality, analyzing them through a two-dimensional framework of "policy tools - policy objectives" to dissect structural characteristics and optimization pathways. Findings show: (1) Imbalanced policy instrument structure: Supply-side (49.74%) and environmental instruments (46.56%) dominate, while demand-side instruments account for merely 3.70%, resulting in weak market pull; (2) Overemphasis on technology transfer at the expense of industrialization: Policies disproportionately focus on technology transfer and application (69.31%), with weak support for industrialization stage (4.76%); (3) Insufficient adaptability between policy tools and transformation stages: The R&D phase neglects quality enhancement (“quality characteristics of outcomes” instruments: 7.14%), while industrialization relies excessively on short-term “funding support” without long-term fiscal and taxation mechanisms; (4) Absence of implementation rules weakens policy effectiveness. Recommendations include strengthening demand-side instruments, balancing three-stage policy allocation, establishing a dynamic “stage instrument” adaptation mechanism, and refining supporting rules to improve policy effectiveness.
  • YANG Guanghui
    High-Technology and Commercialization. 2026, 32(2): 86.
    Abstract ( )   Knowledge map   Save
    This study analyzes the current research status and existing issues in logistics cost management among domestic and international enterprises, summarizing the composition and influencing factors of logistics costs. Leveraging the foundation of “Internet plus,” this research conducts an analysis and study on enterprise logistics costs by optimizing the supply chain model for Company YX. It proposes optimized supply chain logistics models based on activity-based cost indicators and joint distribution, designing four optimization strategies: logistics cost control at nodes, optimized transportation routes, management pathways, and supply chain coordination. The optimization and implementation focus on logistics transportation routes. This provides a reference for other enterprises in managing logistics costs. 
  • SHI Lei
    High-Technology and Commercialization. 2026, 32(2): 89.
    Abstract ( )   Knowledge map   Save
    Human resource management serves as the core backbone of enterprise development, playing a pivotal role in enhancing organizational effectiveness. This paper aims to delve into the intrinsic correlation and optimization pathways between performance evaluation and incentive mechanisms within enterprise human resource management. By analyzing the consistency of their objectives, dynamic feedback loops, and strategic transmission effects, it dissects prevalent issues such as indicator disconnection and feedback deficiency. Subsequently, it proposes optimization measures, including anchoring strategic decoding indicators and constructing a dynamic feedback mechanism. The study reveals that the synergistic operation of the two can achieve resonance between enterprise and employee development, leading to the conclusion that enhancing their effectiveness requires strategic anchoring, dynamic feedback, equitable alignment, and digital empowerment.
  • HAN Bingyi, LI Xiao
    High-Technology and Commercialization. 2026, 32(2): 92.
    Abstract ( )   Knowledge map   Save
    This study examines the challenges in cultivating high-skilled talent for the rail transit sector amid the intelligent transformation. Through questionnaire surveys targeting both enterprises and students, it reveals that rapid technological advancements are exacerbating structural mismatches in talent competencies. A systematic deviation exists between the perceptions of learners and actual market demands, while current teaching and assessment systems fail to effectively support the development of new skill sets. The study proposes establishing a dynamic curriculum system grounded in digital literacy and interdisciplinary integration, implementing educational reforms through a “course-certificate integration and scenario-empowered” training model, and building a comprehensive evaluation ecosystem oriented by competency development and process assessment.
  • ZHOU Weidong, YANG Lu, YANG Fan, LI Yuanyuan, JIANG Dan
    High-Technology and Commercialization. 2026, 32(2): 96.
    Abstract ( )   Knowledge map   Save
    In today’s knowledge economy era, the transformation of scientific and technological achievements is an important driving force for promoting economic growth and social progress and intellectual property protection is an indispensable safeguard mechanism in this process. This paper focuses on the collaborative mechanism between intellectual property protection and the transformation of scientific and technological achievements, and explores its role in promoting scientific and technological innovation, economic development, and social progress. The study shows that intellectual property protection provides a solid legal and institutional foundation for the transformation of scientific and achievements by clarifying the property rights of scientific and technological achievements, incentivizing the enthusiasm of innovative entities, and enhancing the market value of technological achievements. However, the current collaborative mechanism still faces numerous challenges. This paper proposes a series of optimization strategies, providing a theoretical foundation and practical pathways for the implementation of an efficient collaborative mechanism for intellectual property protection and transformation of scientific and technological achievements. These strategies hold practical value for propelling technological breakthroughs and socio-economic advancement.
  • ZHU Bing
    High-Technology and Commercialization. 2026, 32(2): 100.
    Abstract ( )   Knowledge map   Save
    The microbial industry, a key component of the strategic emerging sector of genetic technology, is vital for fostering new growth drivers and building global competitive advantages in China. With the world’s leading number of patent applications for deposited microorganisms, challenges in intellectual property protection arise due to their living nature and technical complexity, including safety compliance, novelty assessment, inventiveness evaluation, and sufficient disclosure in specifications. Based on China’s patent legal framework and practical cases, this study proposes strategies for cross-regulatory safety assessment, prior art evaluation, inventiveness through technical effects, and genetic resource disclosure. These provide standardized guidance for examiners and assist applicants in optimizing patent strategies, promoting the transformation of high-value patents and driving innovation in the microbial industry.
  • KUANG Shuqing, JIANG Tinglan, XU Li
    High-Technology and Commercialization. 2026, 32(2): 104.
    Abstract ( )   Knowledge map   Save
    This study focuses on the application of intellectual property (IP) analysis and evaluation in the full-cycle management of scientific research projects. It analyzes its efficiency mechanism during project initiation, R&D, conclusion, and commercialization phases, clarifying key analysis modules and implementation pathways for each stage. Using a radionuclide separation and extraction technology project as a case study, the paper demonstrates the practical workflow from technical decomposition to decision-making suggestions. The research provides IP analysis and evaluation strategies combining theoretical depth with practical implementation for full-cycle project management. These strategies aim to infuse projects with an IP-conscious mindset within research institutions and maximize innovation value. 
  • HE Junhui
    High-Technology and Commercialization. 2026, 32(2): 110.
    Abstract ( )   Knowledge map   Save
    Due to the particularity of generative artificial intelligence infringement, accurately determining the type of infringement liability in the current attribution principles requires first determining its legal positioning. The legal positioning of generative artificial intelligence as a service form rather than a product form exists because it does not meet the definition of “product” in traditional product liability, and the damage caused by its output information usually does not directly threaten personal safety. Therefore, strict product liability rules should not be applied. This definition is beneficial for encouraging technological innovation and avoiding the unfair phenomenon of users intentionally inducing infringement and demanding developers to take responsibility. To achieve fairness and scientificity in the determination of responsibility, it is recommended to adopt a mechanism of multiple differentiation of responsibility subjects, and allocate responsibility based on the specific roles of technology developers, operators, and users in the infringement chain. Especially for the determination of the fault of service providers, an objective standard of duty of care should be adopted. By applying the responsibility framework of network service providers through analogy and introducing dynamic attribution principles and risk diversification mechanisms, it can provide institutional space for the healthy development of the generative artificial intelligence industry while safeguarding the rights and interests of victims.
  • FU Qiguo, CAO Kun, HUANG Hui
    High-Technology and Commercialization. 2026, 32(2): 115.
    Abstract ( )   Knowledge map   Save
    The chip industry serves as a strategic pillar for the digital economy and national security, and trade secrets play a crucial role in the protection of its implicit technologies. This study points out that trade secrets are not only a key means for chip enterprises to achieve technological breakthroughs and differentiated competition, but also an important driving force for the accelerated upgrading and leapfrog development of the chip industry. Currently, the protection of trade secrets in chip technology faces such dilemmas as inadequate confidentiality measures in enterprises, imperfect judicial authentication mechanisms, and insufficient effectiveness of administrative protection. The study proposes optimized paths, including establishing a full-process confidentiality management system for enterprises, improving the judicial authentication mechanism, and enhancing the effectiveness of administrative protection. It aims to provide theoretical and practical references for building a trade secret protection system that adapts to the development of the chip industry.
  • ZHANG Cancan
    High-Technology and Commercialization. 2026, 32(2): 119.
    Abstract ( )   Knowledge map   Save
    Against the backdrop of the digital economy, intellectual property has become the core competitiveness of small and medium-sized enterprises (SMEs). As an in-depth integration of emerging technologies and traditional financial management, intelligent accounting provides new solutions for SME intellectual property protection. This paper first expounds the significance of intelligent accounting for the protection, then analyzes the prominent problems faced by SMEs, such as imperfect accounting and evaluation systems, low informatization level and weak protection awareness. On this basis, it proposes specific countermeasures including building an intelligent accounting-driven intellectual property asset management system, conducting risk early warning and monitoring via big data analysis, and establishing a full-life-cycle intelligent financial management mechanism, aiming to offer theoretical and practical references for SMEs’ intellectual property protection.
  • YAN Qiong
    High-Technology and Commercialization. 2026, 32(2): 122.
    Abstract ( )   Knowledge map   Save
    With the acceleration of China’s urbanization and the growing demand for underground space development, TBM, as a core equipment for tunnel construction, has seen its technological innovation and intellectual property protection become focal points of industry competition. Based on case studies involving patent information analysis, natural language processing, and knowledge graph construction, this paper systematically explores the integration mechanisms of information construction, intellectual property layout, and knowledge management in the TBM field. The research indicates that: (1) patent intelligence serves as a core basis for assessing technological competition trends and planning research and development directions; (2) knowledge graph technology enables the digital reconstruction of unstructured maintenance knowledge, supporting intelligent decision-making; and (3) the patent layout of intelligent sensing technology highlights technological hotspots and gaps. Finally, a comprehensive framework for intellectual property and knowledge management in TBMs, centered on the theme of “data-driven, knowledge-integrated, intelligent application,” is proposed to provide theoretical references for technological innovation and strategic planning in the industry.  
  • WANG Xiaohua
    High-Technology and Commercialization. 2026, 32(2): 125.
    Abstract ( )   Knowledge map   Save
    With the rapid development of large model technology, training data has become a fundamental element for the performance improvement and capability emergence of generative artificial intelligence. However, the high complexity of training data in terms of sources, structures and usage methods has continuously triggered infringement risks in personal information protection, copyright protection and data rights allocation during the model training and content generation processes. Existing legal norms are mostly based on the assumption of traditional data processing and content production models, making it difficult to effectively respond to the new risks caused by training data for large models. Therefore, starting from the source types and legal attributes of training data, this paper systematically sorts out the main infringement risks in the full life cycle of training data for large models, analyzes the institutional causes, and on this basis, puts forward a legal governance path centered on classified governance, risk orientation and liability allocation, so as to provide normative support for the compliant development of large model technology.