| 张静,谭春林,陈咏梅.学术研究与出版中AIGC伦理失范的表现及其识别策略.编辑学报,2025,37(5):517-523 |
| 学术研究与出版中AIGC伦理失范的表现及其识别策略 |
| Manifestations of ethical misconduct involving AIGC in academic research and publishing, and strategies for its identification |
| |
| DOI：10.16811/j.cnki.1001-4314.2025.05.009 |
| 中文关键词: AIGC伦理 学术诚信 学术失范 出版规范 “标准—技术—人工”识别框架 |
| 英文关键词: AIGC ethics academic integrity academic misconduct publishing standards “Standards–Technology–Human Judgment” framework First-author’s address\ Editorial Office of Journal of Beijing University of Chemical Technology |
| 基金项目:*北京高教学会社会科学学报研究分会2024年度编辑学研究课题(SKXB-YB202405)；广东省科技计划项目—广东省科技期刊优秀人才项目(2024B1212110008)；广东省科技计划项目—广东省科技期刊优秀人才项目(2025A1212150028) |
|
| 摘要点击次数: 1797 |
| 全文下载次数: 617 |
| 中文摘要: |
| 生成式人工智能(GAI)技术的广泛应用对学术诚信体系形成了新的挑战,当前学术研究与出版中的人工智能生成内容(AIGC)伦理失范问题凸显,厘清AIGC不同应用场景、不同责任主体的伦理失范表现形式并探讨其识别策略具有必要性。本研究参照权威的AIGC使用边界指南和伦理准则,分析了从学术研究到学术出版全流程中不同应用场景下相关责任主体可能发生的主观AIGC伦理失范行为,以及大语言模型固有缺陷导致的客观伦理风险。提出了构建“标准—技术—人工”三维识别框架的策略:建议管理部门及时修订增补AI出版行业标准,明确界定AIGC的相关不端行为；合理利用AIGC检测技术辅助识别AI生成内容；发挥人工判断的专业作用,弥补检测工具的不足,倡导通过人机综合判断来提升识别AIGC失范的准确性。 |
| 英文摘要: |
| The swift advancement and extensive deployment of generative artificial intelligence(GAI)technologies have introduced fresh challenges to the academic integrity framework. Ethical transgressions involving artificial intelligence generatecl content(AIGC) in the realms of academic research and publishing have emerged as increasingly conspicuous issues. Hence, it is of paramount importance to elucidate the forms of ethical violations across diverse application scenarios and among various responsible parties, as well as to delve into corresponding identification strategies. By drawing on authoritative guidelines for AIGC utilization, this study scrutinizes both subjective ethical misconduct committed by different stakeholders throughout the academic research-to-publication continuum and objective ethical risks stemming from inherent flaws within AIGC models. The paper puts forward a strategy aimed at establishing a three-dimensional identification framework that integrates standards, technology, and human discernment. Specifically, it calls for regulatory bodies to promptly update publishing standards related to AIGC, to precisely define misconduct associated with AIGC, to judiciously employ AIGC detection technologies to aid in content identification, and to bolster expert human judgment to offset the limitations of automated tools. A collaborative approach involving both human and machine evaluation is advocated to enhance the precision of identifying ethical violations linked to AIGC. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
|
|
