首页|论人工智能生成内容的可版权性标准

论人工智能生成内容的可版权性标准

扫码查看
生成式人工智能将著作权法在独创性表达之上创设的稀缺性和激励转移至表达背后的智力投入,即提示词的迭代中,人工智能生成内容由此具备可版权性基础.现有反对意见过度关注作者身份问题,造成诸多论证困境."意志论"未能区分作者的范畴意图和语义意图,忽视了所有作品的表达内容均不可能在绝对意义上为作者的自由意志所决定、预见和控制,且诉诸"创作辅助人"条款将赋予人工智能法律主体资格,无法成立.而"过程论"将作品可版权性与具体创作方法联系在一起,不符合著作权法的规则和目的.人工智能生成内容的可版权性标准应该从作者身份转向独创性.由于排他权范围极其有限,对人工智能生成内容适用最低限度独创性理论不会造成权利"井喷"现象.
Generative AI shifts the scarcity and incentive created by copyright law from the originality of expression to the intellectual efforts underlying expression,i.e.,the iteration of prompts.This provides a basis for the copyrightability of Generative AI works.The current opposing views overly focus on the issue of authorship,resulting in numerous argumentative challenges.The"will theory"fails to distinguish between the categorical intentions and semantic intentions of the author,overlooking the fact that the expressive content of all works cannot be determined,predicted,or controlled in an absolute sense by the author's free will.Additionally,resorting to the"creative assistant"clauses would confer legal personality on AI,and is not viable.The"process theory"links the copyrightability of a work to specific creative methods,which is inconsistent with the rules and purpose of copyright law.The criterion for the copyrightability of Generative AI works should shift from authorship to originality.Due to the extremely limited scope of exclusive rights,applying the"minimal degree of creativity"to Generative AI works will not lead to a proliferation of rights.

artificial intelligencescarcityauthorshipfree willoriginality

李扬、涂藤

展开 >

中山大学法学院

人工智能 稀缺性 作者身份 自由意志 独创性

国家社会科学基金

18BFX163

2024

知识产权
中国知识产权研究会

知识产权

CSSCICHSSCD北大核心
影响因子:1.115
ISSN:1003-0476
年,卷(期):2024.(1)
  • 71