Optimization of Journal Resource Integration and Dissemination Models from the Perspective of Media Convergence

Jiao Wu
Journal Press, Hebei University, Baoding, Hebei 071002

Abstract

[Objective] The integration of artificial intelligence and journal publishing represents an inevitable choice for the development of information technology and culture. From the perspective of media convergence, journal media face challenges such as difficulties in resource integration and weak dissemination awareness. This paper proposes optimization strategies for AI-enabled journal resource integration and dissemination models.

[Method] First, a journal media intelligent dissemination data system based on the data middle platform concept performs cleaning, transformation, and integration of various types of journal media data to construct a journal dissemination data warehouse. Second, multi-source tagging is employed for user profiling and personalized recommendation of journal users, providing effective support for precise journal media dissemination. Finally, large language model-assisted content generation and editorial optimization are utilized to implement tasks such as overview generation, editing and proofreading, intelligent interaction, and dissemination effect analysis, offering references for journal improvement.

[Results] The optimization of journal dissemination models through artificial intelligence can help journals better adapt to media convergence trends and enhance their dissemination effectiveness and influence.

[Conclusion] As the process of journal media convergence accelerates, intelligent journal dissemination systems have ushered the journal industry into a new stage of practical innovation beyond theoretical exploration. AI-related technologies can effectively promote the reform and optimization of journal dissemination.

Keywords: media convergence; journal publishing; artificial intelligence; scientific journals; dissemination models

Classification Code: G203
Document Code: A
Article ID: 1671-0134(2025)02-94-04
DOI: 10.19483/j.cnki.11-4653/n.2025.02.018

Citation Format: Wu J. Optimization of journal resource integration and dissemination models from the perspective of media convergence[J]. China Media Technology, 2025, 32(2): 94-97.

Introduction

In March 2016, the computer program "AlphaGo" defeated world Go champion Lee Sedol, marking a new stage in the development of Artificial Intelligence (AI). In July 2017, the State Council promulgated the "New Generation Artificial Intelligence Development Plan," and AI began to enter the journal publishing industry and develop rapidly. In August 2019, four departments including the China Association for Science and Technology jointly issued the "Opinions on Deepening Reform and Cultivating World-Class Scientific and Technological Journals," emphasizing the importance of deepening reform of scientific journals and improving their quality and international influence. The document called for seizing the major opportunities of digitalization and intelligence to promote transformation in journal publishing, innovating dissemination mechanisms, and driving the digital transformation and upgrading of scientific journals. In June 2021, three ministries including the Publicity Department jointly issued the "Opinions on Promoting the Prosperous Development of Academic Journals," emphasizing the need to adapt to media convergence trends, adhere to integrated development, achieve full-chain digital transformation and upgrading through process optimization and platform reconstruction, and guide academic journals to adapt to mobile and intelligent development directions by promoting the construction of integrated development platforms. The integration of AI and journal publishing has become an inevitable choice and trend for the development of information technology and culture [1].

Since 2017, domestic academia has conducted preliminary explorations of AI applications in academic journals, with existing research primarily examining AI's impact on journal development, associated problems and challenges, and changes in dissemination models. First, the impact of AI on journal development constitutes the mainstream research direction, with most studies exploring AI applications from the perspective of the journal field. Xu Lifang et al. [2] argue that technology is the catalyst and key force for industry transformation, and that cutting-edge technologies such as AI and blockchain technology, when implemented and applied in the publishing industry, bring new possibilities and directions for journal publishing. An Qi [3] believes that breakthroughs in big data technology and AI technology make it possible to intelligently analyze hot spots and trends in various disciplines and identify outstanding author resources in these areas. Lü Xuemei et al. [4] propose that Chinese scientific and technological journal publishing institutions can improve the accuracy of online translation systems and enhance the international visibility of Chinese scientific journals by leveraging AI and developing parallel corpora and Chinese reprocessing methods. Liu Chang et al. [5] argue that AI can not only improve the efficiency and accuracy of topic planning but also enhance the matching degree of solicited contributions.

Second, some scholars have noted that AI may bring new problems and challenges to journal development. Dai Ni et al. [6] suggest that while AI accelerates the speed of scientific publishing, it may cause scientific literature proliferation, lose the filtering function of scientific publishing, and make it difficult to ensure the quality of published papers. Pan Xue et al. [7] believe that intelligent publishing faces many practical dilemmas, including conflicts between urgent needs and scarce technical resources, conflicts between creative dissemination and copyright risks, and conflicts between precise push and algorithmic discrimination. Liu Jiangxia et al. [8] argue that scientific journal editors face challenges such as the reconstruction of digital platforms, editorial and publishing work content, information terminal systems, and journal presentation forms. They must reshape core competencies including professional academic literacy, new technology application literacy, organizational management ability, service awareness, and scientific and technological ethics literacy to comprehensively improve their ability to cope with intelligent media development.

Third, a few scholars have focused on the important influence of AI on journal dissemination. Zheng Quan [9] believes that under the situation of media convergence, we should explore ways to improve the precise dissemination ability of scientific and technological journals by expanding academic search paths, building personalized precise push platforms and diversified dissemination models, and providing targeted services to users. Huang Ying et al. [10] discuss the application and prospects of mainstream AI technologies, including natural language processing, data mining, intelligent recommendation, machine learning, and AI writing, in journal dissemination.

1. Current Status of Journal Resources and Dissemination from the Perspective of Media Convergence

The journal dissemination model refers to the channels and methods through which journal content is transmitted from editors to readers under different periods and technological environments, encompassing a series of processes including journal content integration, processing, publishing, and dissemination. This model reflects how journals convey their content to target audiences and enhance their influence and coverage through different channels and technological means. With the continuous iteration of media, academic research achievements can no longer be limited to printed academic journals. Various experimental data, audio, video, and manuscripts have appeared extensively in both formal and informal academic publishing venues. Social media platforms have become venues for vibrant debates on fresh academic viewpoints, attracting a broader and more diverse media audience and forming new academic dissemination networks. WeChat groups and official accounts composed of different academic institutions or communities have become the most popular academic publishing venues beyond printed journals. Online databases, WeChat official accounts, and journal official websites have become the main channels for audiences to access academic literature.

The proliferation of technology, particularly the infiltration of AI technology applications, is profoundly transforming journal dissemination models. Examples include intelligent editing and review systems, intelligent recommendation services, and knowledge graph construction, making journal dissemination more efficient, precise, and interactive. There have been some successful cases both domestically and internationally: the open-access journal publisher Frontiers provides AI software that helps editors, reviewers, and authors automatically evaluate the quality of academic papers, provide revision suggestions, and improve publishing efficiency [6]; TrendMD's content recommendation engine can push journal manuscripts to multiple academic sites, automatically recommending manuscripts through data mining algorithms and delivering content to potentially interested readers, achieving precise dissemination and intelligent sharing [6]; the Chinese Journal of Endocrine Surgery has built its own database based on articles since its inception, using big data mining and precise matching technology to collect information on authors, institutions, and articles, forming different reader and author groups by discipline, and achieving precise paper push services by sending emails to target recipients [7].

Due to weak dissemination awareness, slow updating of publishing technology, and lack of media literacy among practitioners, AI technology has not yet been truly integrated into domestic academic journal publishing. Changing this situation requires more resource investment and the establishment of an intelligent journal media dissemination platform.

2. Optimization Strategies for Journal Resource Integration and Dissemination Models

2.1 Journal Media Intelligent Dissemination Data System Based on the Data Middle Platform Concept

With the deepening advancement of journal media convergence, media data from third-party journal platforms, journal sites, and social networking platforms are becoming increasingly valuable for journal dissemination, making effective integration and utilization of journal media data highly significant. The "data middle platform" represents a new data management concept proposed from the perspective of data assets. The "journal data middle platform" uses technical means to collect, integrate, store, and process multi-source heterogeneous journal media data while unifying data standards to form a journal big data asset layer, providing efficient data services for internal and external AI applications in journal dissemination [11-12]. By establishing a data integration and utilization framework, cross-platform data distributed collection and structured extraction of heterogeneous data can be achieved. Big data technology is used to process large amounts of raw data from social media through deduplication and aggregation, metadata definition, type conversion, association analysis, and text mining, completing the cleaning, transformation, and integration of various types of journal media data to construct a journal dissemination data warehouse.

This paper refers to data middle platform construction methodology [13], designs a "journal media intelligent dissemination data system" through essential analysis of journal media resources and dissemination models and induction of journal media data collections. This system can be divided from bottom to top into the source data layer, entity fusion layer, relationship fusion layer, profiling tag layer, intelligent algorithm layer, and dissemination application layer. The source data layer is responsible for collecting original journal knowledge data from third-party journal platforms, journal sites, and social networking platforms to establish a journal media knowledge base; the entity fusion layer completes entity alignment across different knowledge bases and associates entities from different data domains; the relationship fusion layer uses translation models to merge key equivalent relationships from different journal media knowledge bases; the profiling tag layer performs user profiling for journal media users and generates user tag sets; the intelligent algorithm layer covers algorithm models and parameter sets for different purposes, providing algorithmic support for intelligent dissemination; and the dissemination application layer includes various journal media applications to achieve intelligent dissemination across media matrices.

2.2 Journal User Profiling and Personalized Recommendation Based on Multi-Source Tags

Precise dissemination represents an important method in journal media communication, and profiling journal media users constitutes a fundamental task for precise dissemination [14]. Journal user profiling for multi-source journal media includes the following steps [15]: First, user attribute tag generation. Based on the multi-source journal media knowledge base from the source data layer, journal user information is collected and cleaned to generate user attribute tags, focusing on basic attribute information including name, Open Researcher and Contributor ID (ORCID), work unit, position and title, birth date, native place, research field, and email address, some of which serve as unique user identifiers. Second, user interest tag mining. By analyzing user behavior records across journal media (such as browsing, downloading, commenting, and liking journal articles), methods including TF-IDF, LDA topic models, natural language models, and word vector mining are employed to calculate users' interested disciplines, subjects, or keywords and mine their interest tag sets. Third, user interest tag expansion. User tags generated in the first two steps may suffer from data sparsity and cold-start problems. To enrich user tags, collaborative filtering thinking can be used to expand user tags—similar users often have similar interests. For example, if existing user tag data shows that users who like "artificial intelligence" also like "big data," then the "big data" tag can be extended to users with the "artificial intelligence" tag. Fourth, cross-domain interest tag fusion. Different journal media belong to different data domains, and fusing interest tags from multi-domain media can more effectively depict user interests. User interest ranking can also be achieved through tag weighting [12].

Tags stored in the profiling tag layer are mainly divided into two types: journal user tags and journal content tags. User tags serve user profiling to represent user interest characteristics; journal content tags are extracted from the journal media content knowledge base to represent journal article content features. By using intelligent algorithms to match and predict user tags with content tags, personalized recommendations for journal users can be achieved. Currently, industry recommendation algorithms for knowledge fusion mainly include embedding-based recommendation models, link relationship-based recommendation models, and propagation-based recommendation models. These recommendation models can leverage mainstream AI algorithms to precisely push journal media article content to journal users.

2.3 Large Language Model-Assisted Content Generation and Editorial Optimization

Large language models have become powerful tools for scientific journals to improve content generation efficiency and editorial optimization [16], with relevant improvements主要体现在 three aspects. First, new media article overview generation. In journal new media dissemination, paper abstracts contain too little information while full texts are too lengthy, which does not align with users' fast-reading habits. Large language models can automatically generate article overviews suitable for online dissemination based on original papers, publishing concise versions through new media platforms such as social networks and short videos to provide audiences with a new reading experience. Second, new media content editing and proofreading. In journal new media content publishing and dissemination, article proofreading and style adjustment are time-consuming and tedious. Large language models can automatically detect grammatical errors in text and provide improvement suggestions while optimizing content style to ensure standardization and consistency across various media platforms. Third, enhanced intelligent interaction with readers. By building a language model-based question-answering system, journal media can provide a platform where readers can ask questions about article content and receive instant, text-based answers, enhancing reader interaction. For complex scientific concepts, video models can assist in creating AR/VR content to help readers understand these concepts interactively, providing immersive learning experiences. Fourth, cross-platform dissemination effect analysis. Through large language model agents analyzing content dissemination effects across different journal media platforms, such as click-through rates, sharing rates, and engagement levels, editors are provided with data support to help optimize dissemination strategies. By analyzing search engine algorithms and user search habits, journals can be helped to optimize article titles, keywords, and descriptions to improve search engine rankings.

3. Challenges and Development of AI in Journal Dissemination

Although AI applications in journal dissemination provide many conveniences and innovative possibilities, they also face a series of challenges. First, accuracy and reliability issues. AI-generated content may contain errors or misleading information, particularly when handling complex scientific and technological topics where biases or even common-sense problems can easily arise. The output quality of AI models highly depends on the quality of input data, and inaccurate or biased data can lead to misleading results. Second, legal and ethical issues. AI-generated content may involve copyright issues, especially when existing literature, images, and data are used, making content sources difficult to trace. When generated content leads to misunderstandings or other problems, determining responsibility attribution can be complicated. When using user data for personalized recommendations, it is also necessary to ensure compliance with relevant privacy protection regulations. Third, technical and resource issues. Developing and maintaining advanced AI-assisted systems requires significant investment, mainly including computing resources, algorithm models, and personnel capabilities. Additionally, journal sponsors often lack relevant technical capabilities in the process of developing, training, and managing models. Fourth, algorithm transparency and interpretability issues. Due to the black-box working mode of AI, the decision-making process of algorithm models is opaque and output content is uncertain, leading to trust issues for the journal industry with strict content management requirements, especially when needing to explain how AI reaches specific conclusions [17].

To better promote the development of AI in the field of journal dissemination, future efforts will maximize AI's development potential from multiple perspectives. AI systems will adopt more advanced algorithms and larger-scale datasets to improve content generation accuracy, better understanding and processing complex scientific topics through continuous advances in machine learning and natural language processing. Meanwhile, journal publishers may gradually establish stricter data quality control mechanisms to ensure that training data for AI models is accurate, comprehensive, and unbiased. Copyright issues will be resolved through innovative authorization mechanisms and intelligent copyright management systems, with responsibility attribution and privacy protection addressed by establishing clear responsibility frameworks and strengthening algorithm audits. The development cost of AI may decrease with technological maturation and large-scale application, with open-source algorithm models and cloud services for the journal industry enabling journal sponsors to obtain required computing resources and algorithm models at lower costs. Future AI models will place greater emphasis on transparency and interpretability, using visualization technologies such as knowledge graphs to enable users to intuitively understand AI decision-making processes.

4. Conclusion and Outlook

This paper analyzes the development status of the journal field in the era of intelligent media and proposes optimization strategies for AI-enabled journal dissemination models in response to changes in journal dissemination patterns from the perspective of media convergence. First, a journal media intelligent dissemination data system based on the data middle platform concept cleans, transforms, and integrates various types of journal media data to construct a journal dissemination data warehouse. Second, multi-source tagging is used for journal user profiling and personalized recommendation, providing effective support for precise journal media dissemination. Third, large language model-assisted content generation and editorial optimization implement tasks such as overview generation, editing and proofreading, intelligent interaction, and dissemination effect analysis, providing references for journal improvement. AI is not merely an auxiliary tool but will become an indispensable part of the journal dissemination field. Through the optimization of journal dissemination models, AI can help journals better adapt to media convergence trends and enhance their dissemination effectiveness and influence.

References

[1] Xue Chunlu, Wang Yuanjie, Liu Jifang, et al. Optimization of smart publishing and dissemination models for scientific journals based on intelligent new technologies[J]. Acta Editologica, 2023(S1).

[2] Xu Lifang, Wang Yu. Innovation in the scientific content industry: Research on overseas scientific journal publishing trends in 2018[J]. Science-Technology & Publication, 2019(2): 13-20.

[3] An Qi. New explorations in academic journal development in the big data era[J]. Chinese Editors, 2017(7): 57-61.

[4] Lü Xuemei, Cheng Lidong, Zhang Hong, Cheng Jianxia. Enhancing the international visibility of Chinese scientific journals through online translation systems[J]. Chinese Journal of Scientific and Technical Periodicals, 2019(2): 173-178.

[5] Liu Chang, Jiang Jingmei, Fan Yuxian. Application and coping strategies of artificial intelligence in topic planning for scientific journals[J]. Chinese Journal of Scientific and Technical Periodicals, 2020(8): 34-38.

[6] Dai Ni, Bu Zhaode. The impact of artificial intelligence on scientific journal publishing[J]. Publishing Wide Angle, 2021(11): 46-48, 51.

[7] Pan Xue, Zhang Haisheng, Guo Lei. Development prospects, practical dilemmas, and promotion strategies of intelligent publishing for scientific journals[J]. Acta Editologica, 2022(4): 378-383.

[8] Liu Jiangxia, Zhu Ying. Challenges and countermeasures for scientific journal editors in the era of intelligent media[J]. Acta Editologica, 2021(6): 620-624.

[9] Zheng Quan. Countermeasures and reflections on improving the precise dissemination capability of scientific journals in the era of media convergence[J]. Acta Editologica, 2020(2): 188-190.

[10] Huang Ying, Yang Hao, Wang Daoyou. Application and prospects of artificial intelligence in scientific journal dissemination[J]. Acta Editologica, 2023(S1): 137-140.

[11] Shen Xibin, Ma Ming, Liu Hongxia, et al. Strategic research and practice on building an integrated academic service platform for scientific journals based on digital middle platform[J]. Acta Editologica, 2022(3): 306-311.

[12] Zhang Bin. Research on cross-domain recommendation based on knowledge graph[D]. Baoding: Hebei University, 2023: 152.

[13] Fu Dengpo, Jiang Min, Ren Yinzi, et al. Data Middle Platform: Making Data Work[M]. Beijing: China Machine Press, 2021.

[14] Li Baopeng, Liu Bin. Construction of a scientific journal dissemination system based on precise dissemination methods[J]. Communication and Copyright, 2024(2): 69-71.

[15] Wang Yajiao, Lu Jia, Ke Xiaojing. Research on the application of academic profiling in scientific journals[J]. Chinese Editors, 2021(4): 45-49.

[16] Xu Jinghong, Zhang Rukun. Risks and coping strategies of large language model applications in academic publishing[J]. Chinese Editors, 2024(2): 36-42.

[17] Song Shilei, Yang Yiyun. Application scenarios, risks, and prospects: Academic publishing in the era of ChatGPT-like large language models[J]. Publishing Science, 2023(5): 23-30.

Funding: Hebei Provincial Social Science Fund Project "Research on Data Resource Integration and Influence Enhancement of University Humanities Journals in the New Media Convergence Era" (Project No.: HB22TQ004).

Author Biography: Wu Jiao (1980—), female, from Wuhan, Hubei, holds a master's degree and is an associate editor. Her research focuses on editing and publishing.

(Responsible Editor: Li Yansong)