Scholarlink

<p>Artificial intelligence (AI) research has grown rapidly in recent years, reflected in the substantial increase of AI-related publications indexed in major databases. This rapid expansion has intensified the demand for timely dissemination mechanisms, highlighting the importance of Early Access (EA) publication as a channel for accelerating knowledge exchange. However, limited research has examined the readership characteristics and dissemination patterns of AI articles published as Early Access (AI-EA). This study investigates the academic roles and disciplinary backgrounds of AI-EA readers, examines role-based and disciplinary reader networks, and analyzes the relationship between reader counts and the citation performance of these articles. We constructed a dataset of 3364 AI-EA articles published between 2024 and 2025 in the Web of Science Core Collection. Using the Mendeley API, we retrieved metadata and reader statistics through DOI matching, achieving a 93% coverage rate. In total, 3128 articles attracted 16,652 readers, with demographic data available for 7487 of them. PhD (21.09%) and Master (14.41%) led the student group, while researchers (20.57%) and lecturers (10.47%) dominated the faculty readership. Disciplinary analysis shows that most readers were from computer science (42.49%) and engineering (19.43%), with significant engagement also observed from the social sciences (17.80%), life sciences (7.59%), natural sciences (5.76%), and even arts and humanities (4.55%). Co-occurrence network reveals strong clustering among technical disciplines, alongside meaningful interdisciplinary links. Role-based network analysis identifies graduate students, researchers, and lecturers as key diffusion nodes. A weak but statistically significant Spearman correlation (<i>ρ</i> = 0.209, <i>p</i> &lt; 0.001) between reader counts and citations suggests that early readership may provide predictive signals of future academic impact. These findings offer new insights into early-stage scholarly interaction with AI-EA and inform more targeted dissemination strategies.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

First to know: profiling readers of early-access articles in the field of AI

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月30日

查看全文
  • Yi-Shuai Xu,
  • A. M. K. Yanti Idaya,
  • Muhammad Shahreeza Safiruz Kassim

摘要

Artificial intelligence (AI) research has grown rapidly in recent years, reflected in the substantial increase of AI-related publications indexed in major databases. This rapid expansion has intensified the demand for timely dissemination mechanisms, highlighting the importance of Early Access (EA) publication as a channel for accelerating knowledge exchange. However, limited research has examined the readership characteristics and dissemination patterns of AI articles published as Early Access (AI-EA). This study investigates the academic roles and disciplinary backgrounds of AI-EA readers, examines role-based and disciplinary reader networks, and analyzes the relationship between reader counts and the citation performance of these articles. We constructed a dataset of 3364 AI-EA articles published between 2024 and 2025 in the Web of Science Core Collection. Using the Mendeley API, we retrieved metadata and reader statistics through DOI matching, achieving a 93% coverage rate. In total, 3128 articles attracted 16,652 readers, with demographic data available for 7487 of them. PhD (21.09%) and Master (14.41%) led the student group, while researchers (20.57%) and lecturers (10.47%) dominated the faculty readership. Disciplinary analysis shows that most readers were from computer science (42.49%) and engineering (19.43%), with significant engagement also observed from the social sciences (17.80%), life sciences (7.59%), natural sciences (5.76%), and even arts and humanities (4.55%). Co-occurrence network reveals strong clustering among technical disciplines, alongside meaningful interdisciplinary links. Role-based network analysis identifies graduate students, researchers, and lecturers as key diffusion nodes. A weak but statistically significant Spearman correlation (ρ = 0.209, p < 0.001) between reader counts and citations suggests that early readership may provide predictive signals of future academic impact. These findings offer new insights into early-stage scholarly interaction with AI-EA and inform more targeted dissemination strategies.