Scholarlink

Objective <p>The information age has transformed technologies across disciplines. Generative artificial intelligence (GenAI), as an emerging technology, has integrated into scientific research. Recent studies identify GenAI-related scientific research integrity concerns. Using Complex Adaptive Systems (CAS) theory, this research examines risk factors and preventive measures for each agent within the scientific research integrity management system during GenAI adoption, providing new perspectives for integrity management.</p> Method <p>This study applies CAS theory to analyze the scientific research integrity management system, identifying four core micro-level agents: researchers, research subjects, scientific research administrators, and academic publishing institutions. It examines macro-system complexity, agent adaptability, and the impact of agent interactions on the overall system. This framework enables analysis of GenAI’s effects on the research integrity management system.</p> Results <p>The scientific research integrity management system exhibits structural, hierarchical, and multidimensional complexities, with internal circulation of policy, funding, and information elements. In response to GenAI integration, four micro-level agents—researchers, research subjects, scientific research administrators, and academic publishing institutions—adapt their behaviors to systemic changes. Through these interactions, behavioral outcomes emerge at the macro level, driving evolution of the research integrity management system.</p> Conclusions <p>Risks of scientific misconduct permeate the entire research process and require urgent governance. This study recommends that scientific research administrators promptly define applicable boundaries for GenAI in research to guide researchers. Concurrently, they should collaborate with relevant departments to establish regulatory frameworks addressing potential GenAI-related misconduct. Academic publishing institutions must assume quality assurance responsibilities by strengthening review and disclosure protocols. Furthermore, research integrity considerations should be systematically integrated into GenAI’s technological development and refinement.</p> Highlights <p>● Develops an analytical framework grounded in Complex Adaptive Systems (CAS) theory to map evolving interactions among researchers, research subjects, scientific research administrators, and academic publishing institutions within GenAI-integrated research ecosystems.</p> <p>&#xa0;●&#xa0;Identifies self-reinforcing dynamics between GenAI adoption and integrity governance, wherein adaptive rule adjustments by agents reshape system-wide integrity thresholds.</p> <p>&#xa0;●&#xa0;Proposes adaptive governance mechanisms that balance innovation safeguards with integrity guardrails, emphasizing context-sensitive policy calibration over universal solutions.</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

From chaos to symbiosis: exploring adaptive co-evolution strategies for generative AI and research integrity systems

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月26日

查看全文
  • Wenqing Miao,
  • Huan Zang,
  • Qirui Liu,
  • Tianlei Zheng,
  • Yan Zhou,
  • Chunmei Liu,
  • Na Yang,
  • Hengzhi Zhang,
  • Yuwan Zhang,
  • Ying Zhang,
  • Shengli Li,
  • Shenyang Zhang,
  • Hao Zhang

摘要

Objective

The information age has transformed technologies across disciplines. Generative artificial intelligence (GenAI), as an emerging technology, has integrated into scientific research. Recent studies identify GenAI-related scientific research integrity concerns. Using Complex Adaptive Systems (CAS) theory, this research examines risk factors and preventive measures for each agent within the scientific research integrity management system during GenAI adoption, providing new perspectives for integrity management.

Method

This study applies CAS theory to analyze the scientific research integrity management system, identifying four core micro-level agents: researchers, research subjects, scientific research administrators, and academic publishing institutions. It examines macro-system complexity, agent adaptability, and the impact of agent interactions on the overall system. This framework enables analysis of GenAI’s effects on the research integrity management system.

Results

The scientific research integrity management system exhibits structural, hierarchical, and multidimensional complexities, with internal circulation of policy, funding, and information elements. In response to GenAI integration, four micro-level agents—researchers, research subjects, scientific research administrators, and academic publishing institutions—adapt their behaviors to systemic changes. Through these interactions, behavioral outcomes emerge at the macro level, driving evolution of the research integrity management system.

Conclusions

Risks of scientific misconduct permeate the entire research process and require urgent governance. This study recommends that scientific research administrators promptly define applicable boundaries for GenAI in research to guide researchers. Concurrently, they should collaborate with relevant departments to establish regulatory frameworks addressing potential GenAI-related misconduct. Academic publishing institutions must assume quality assurance responsibilities by strengthening review and disclosure protocols. Furthermore, research integrity considerations should be systematically integrated into GenAI’s technological development and refinement.

Highlights

● Develops an analytical framework grounded in Complex Adaptive Systems (CAS) theory to map evolving interactions among researchers, research subjects, scientific research administrators, and academic publishing institutions within GenAI-integrated research ecosystems.

 ● Identifies self-reinforcing dynamics between GenAI adoption and integrity governance, wherein adaptive rule adjustments by agents reshape system-wide integrity thresholds.

 ● Proposes adaptive governance mechanisms that balance innovation safeguards with integrity guardrails, emphasizing context-sensitive policy calibration over universal solutions.