新一代人工智能生成内容档案身份的认定风险及规制研究—基于对ChatGPT生成内容的思考

doi:10.16065/j.cnki.issn1002-1620.2023.05.001

档案学研究

2023, Vol. 37

Issue (5): 4-12 DOI: 10.16065/j.cnki.issn1002-1620.2023.05.001

基础理论研究

本期目录 | 过刊浏览

新一代人工智能生成内容档案身份的认定风险及规制研究—基于对ChatGPT生成内容的思考

陈艳红,李健

湘潭大学公共管理学院湘潭 411105

Research on the Identification Risk and Regulation of the Identity of the Content File of Artificial Intelligence: Thinking About the Content Generated by ChatGPT

CHEN Yanhong,LI Jian

School of Public Administration, Xiangtan University, Xiangtan 411105

全文:

HTML

PDF(8602 KB)
输出: BibTeX | EndNote (RIS)

摘要:

作为新一代人工智能典范，ChatGPT的出现正在给内容生成产业乃至整个信息管理行业带来巨大变革。从档案学专业视角出发，对新一代人工智能生成内容进行“档案身份”的认定成为摆在档案学学者面前亟须论证和解决的重点课题。一方面，ChatGPT技术在优化档案业务流程、深入挖掘档案数据价值、进行档案信息智能检索等方面具有可操作性。但另一方面，ChatGPT生成内容由于高度依赖语料库及算法模型，生成内容本体可能会出现常识性错误、政治偏见等问题。同时，生成内容实体是否属于“作品”、生成内容的“著作归属权”以及“侵权责任”等的认定还面临诸多困境，以上风险和困境加大了对人工智能生成内容档案身份认定的难度。为解决这一问题，可以通过从源头上“控制”内容生成主体、软件开发主体、行业管理主体等来规范生成内容的真实性和可靠性，进而认定人工智能生成内容的档案身份。

关键词 ：人工智能技术, 生成内容, 档案身份, 认定风险

Abstract：

As a model of the new generation of artificial intelligence, the emergence of ChatGPT is bringing great changes to the content generation industry and even the entire information management industry. From the perspective of archival science, the identification of "archival identity" for the content generated by the new generation of artificial intelligence has become a key issue that needs to be demonstrated and solved urgently by archival scholars. On the one hand, ChatGPT technology is operable in terms of optimizing archives business processes, digging deep into the value of archives data, and performing intelligent retrieval of archives information. But on the other hand, because the content generated by ChatGPT is highly dependent on the corpus and algorithm model, the generated content ontology may have common sense errors, political bias, etc. At the same time, there are still many difficulties in determining whether the generated content entity is a "work" "authorship" and "infringement liability" of the generated content. The above risks and difficulties have increased the difficulty of identifying the identity of artificial intelligence-generated content files. In order to solve this problem, we can standardize the authenticity and reliability of the generated content by "controlling" the content generation subject, software development subject, and industry management subject from the source, and then identify the profile identity of the content generated by artificial intelligence.

Key words： artificial intelligence technology generating content archival identity identification risk

出版日期: 2024-10-28

	服务

	把本文推荐给朋友
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	陈艳红
	李健

引用本文:

陈艳红, 李健. 新一代人工智能生成内容档案身份的认定风险及规制研究—基于对ChatGPT生成内容的思考[J]. 档案学研究, 2023, 37(5): 4-12.
CHEN Yanhong, LI Jian. Research on the Identification Risk and Regulation of the Identity of the Content File of Artificial Intelligence: Thinking About the Content Generated by ChatGPT. Archives Science Study, 2023, 37(5): 4-12.

链接本文:

http://journal12.magtechjournal.com/Jwk_dax/CN/10.16065/j.cnki.issn1002-1620.2023.05.001 或 http://journal12.magtechjournal.com/Jwk_dax/CN/Y2023/V37/I5/4

[1]	财联社. ChatGPT打破TikTok最快突破1亿月活产品记录仅用时不到3个月达成[EB/OL]. [2023-03-25]. https://www.163.com/dy/article/HSM0SNFV05198CJN.html.
[2]	彭新. Gartner: 预计2025年超10％的数据将由“生成式AI”创造[EB/OL]. [2023-03-25]. https://new.qq.com/rain/a/20211120A02UWV00l.
[3]	Sonya Huang&Pat Grady. Generative AI: a creative new world[EB/OL]. [2023-03-25]. https://www.sequoiacap.com/article/generative-ai-a-creative-new-world/.
[4]	DIST上海数慧. 史上最快！ChatGPT用户破亿后,又有一个大动作,港股巨头也涨破四千亿[EB/OL]. [2023-03-25]. https://wallstreetcn.com/articles/3681014.
[5]	徐继敏, 严若冰. 论大语言模型材料的证据属性—以ChatGPT和文心一言为例[J/OL]. 四川师范大学学报(社会科学版). https://doi.org/10.13734/j.cnki.1000-5315.2023.0316.
[6]	尹克寒. ChatGPT的发展对情报信息工作的影响及启示[J]. 图书馆理论与实践, 2023(3):15-22.
[7]	吴若航, 茆意宏. ChatGPT热潮下的图书馆服务:理念、机遇与破局[J]. 图书与情报, 2023(2):34-41.
[8]	张海, 刘畅, 王东波, 等. ChatGPT用户使用意愿影响因素研究[J]. 情报理论与实践, 2023(4):15-22.
[9]	宋士杰, 赵宇翔, 朱庆华. 从ELIZA到ChatGPT: 人智交互体验中的AI生成内容(AIGC)可信度评价[J]. 情报资料工作, 2023(4):35-42.
[10]	周文欢. ChatGPT在档案领域应用和意义[J]. 中国档案, 2023(3):62-63.
[11]	Beijaard D. Teachers' prior experiences and actual perceptions of professional identity[J]. Teachers and teaching, 1995(2):281-294.
[12]	张锦. “口述档案”之档案身份及其学理革命契机[J]. 档案学通讯, 2020(3):20-28.
[13]	刘娜. 社会档案意识视野下的档案身份认同策略研究[J]. 湖北档案, 2013(11):18-19.
[14]	黄霄羽, 张斌. 以创新应对变化以发展追求认同—第17届国际档案大会评析[J]. 档案学通讯, 2013(1):4-7.
[15]	徐富荣. 谈电子档案原始性认定问题与对策[J]. 档案学研究, 2001(6):48-50.
[16]	郑满宁. 人工智能技术下的新闻业:嬗变、转向与应对—基于ChatGPT带来的新思考[J]. 中国编辑, 2023(4):35-40.
[17]	李健, 王运彬. 多元主体协同视角下我国参与式档案治理路径研究[J]. 浙江档案, 2022(7):25-28.
[18]	中国信通院. 人工智能生成内容(AIGC)白皮书(2022年)[EB/OL]. [2023-03-25]. http://www.caict.ac.cn/sytj/202209/t20220913_408835.htm.
[19]	陈永伟. 超越ChatGPT: 生成式AI的机遇、风险与挑战[J]. 山东大学学报(哲学社会科学版), 2023(3):127-143.
[20]	雷丽莉, 朱硕. 人工智能生成稿件权利保护问题初探—基于Dreamwriter著作权案的分析[J]. 传媒观察, 2022(5):62-69.
[21]	宋先发, 龚模干. 人工智能生成物著作权问题研究—以ChatGPT生成物为例[J]. 互联网天地, 2023(4):34-41.
[22]	YJango. 万字科普GPT4为何会颠覆现有工作流[EB/OL]. [2023-04-15]. https://www.bilibili.com/video/BV1MY4y1R7EN/?spm_id_from=333.337.search-card.all.click.
[23]	贾骥业, 王林. ChatGPT爆火:伦理安全拷问现行治理体系[N]. 中国青年报,2023-02-21(6).
[24]	许纪霖. 跳出魔盒的精灵:ChatGPT与人类的两难困境—以沉浸式用户体验为例[J]. 探索与争鸣, 2023(3):75-81.
[25]	Friedman B. Bias in computer systems[J]. Human Values & the Design of Computer Technology, 1996(3):330-347.
[26]	Draude C, Klumbyte G, P Lücking, et al. Situated algorithms: a sociotechnical systemic approach to bias[J]. Online Information Review, 2020(2):325-342.
[27]	熊文景. 档案正义论[J]. 档案学通讯, 2021(6):109-112.
[28]	张萌, 朱鸿军. 知识暗流的合规实践:ChatGPT在学术出版中的应用与挑战[J]. 科技与出版, 2023(5):33-40.
[29]	易继明. 人工智能创作物是作品吗?[J]. 法律科学(西北政法大学学报), 2017(5):137-147.
[30]	Rita Matulionyte. With artificial intelligence now generating original poetry, novels and art, Macquarie University's Dr Rita Matulionyte says the law is lagging behind on issues of copyright ownership[EB/OL]. [2023-03-25]. https://lighthouse.mq.edu.au/article/december-2019/AI-generated-art-who-owns-the-copyright.
[31]	Joanna Gill. AI-generated art is booming-but who really owns it?[EB/OL]. [2023-03-25]. https://www.weforum.org/agenda/2022/09/artificial-intelligence-ai-generated-art-ethics-copyright/.
[32]	沈思言, 刘建. 人工智能数据新闻作品著作权归属问题探析[J]. 中国出版, 2019(8):52-55.
[33]	苗成林. 人工智能创作作品的版权归属研究[J]. 科学技术哲学研究, 2021(2):123-128.
[34]	한지영. 인공지능과 법-인공지능 창작물의 권리귀속에 관한 검토[J]. 아주법학, 2022(4):335-370.
[35]	Rodrigues R. Legal and human rights issues of AI: gaps, challenges and vulnerabilities[J]. Journal of Responsible Technology, 2020(4):100005.
[36]	Turner J. Legal personality for AI: regulating artificial intelligence[M]. London: Palgrave Macmillan, 2019:173-205.
[37]	Jaynes T L. Legal personhood for artificial intelligence: citizenship as the exception to the rule[J]. AI & Society, 2020(2):343-354.
[38]	曲春梅. 解构与重建:后现代背景下对档案职业公信力的思考[J]. 档案学通讯, 2016(3):8-12.
[39]	胡吉明, 阳巧英. 认知与保障:档案学语境下的“真实性”研究[J]. 档案学研究, 2021(6):41-46.
[40]	量子位. ChatGPT竟写出毁灭人类计划书,还给出相应Python代码[EB/OL]. [2023-03-25]. https://www.163.com/dy/article/HNQS6S520511DSSR.html.
[41]	Shane Raymond. Deepfake anyone? AI synthetic media tech enters perilous phase[EB/OL]. [2023-03-25]. https://www.reuters.com/technology/deepfake-anyone-ai-synthetic-media-tech-enters-perilous-phase-2021-12-13/.
[42]	新民周刊. “危险”的ChatGPT: 法律、伦理和哲学挑战[EB/OL]. [2023-03-25]. https://new.qq.com/rain/a/20230216A051CC00.
[43]	刘艳红. 生成式人工智能的三大安全风险及法律规制—以ChatGPT为例[J]. 东方法学, 2023(4):29-43.
[44]	新华网. 当人工智能“列车”迎面而来[EB/OL]. [2023-03-25]. https://news.ifeng.com/a/20160315/47878831_0.shtml.

[1]	陆阳, 蔡之玲. 档案与身份认同研究现状考察与进路展望[J]. 档案学研究, 2021, 35(1): 32-39.
[2]	徐拥军, 陈洁, 张晨文, 郭欣雨. “用户生成内容”在档案工作中的应用及影响[J]. 档案学研究, 2019, 33(2): 13-19.
[3]	陶水龙, 薛四新, 田雷, 张桂刚, 李超. 异构系统中电子档案凭证性保障的整体构思[J]. 档案学研究, 2012, 26(5): 50-54.

Viewed

Full text

Abstract

Cited

Shared

Discussed