AI技术介入下胶片影像数字修复的著作权问题与制度重塑研究

Research on Copyright Issues and Institutional Reconstruction of Film Images Digital Restoration Intervened by AI Technology

摘要: 随着生成式人工智能与深度学习算法在文化遗产保护领域的广泛应用，AI修复技术已成为抢救与活化老旧胶片影像的核心驱动力。然而，AI技术的深度介入对以“人类作者”为中心的传统著作权制度带来了根本性冲击。本文立足于我国现行《著作权法》基本原理，首先运用技术类型化方法厘清胶片影像AI修复的技术分层，将其划分为基础级、中度级与深度级三个层级，并分别对应数字化复制、演绎改编与独立创作三种法律行为定性。进言之，本文深度解剖了AI修复成果在“独创性”判定中所遭遇的教义学困境，揭示了“修旧如旧”的审美诉求与著作权法“新意表达”之间的逻辑悖论；系统分析了在算法提供者、技术平台、操作人员与原始版权人多方利益交织下的权利归属纠纷以及公有领域数字垄断风险。最后，结合我国《民法典》与《著作权法》的立法精神，本文提出了针对性的制度应对路径：建构“主客观双轨制”的独创性精细化判定模型；探索引入专门针对胶片数字修复成果的“技术修复权”或邻接权保护机制；完善针对公有领域影像修复成果的合理限制与著作权集体管理机制，从而在鼓励技术投资、尊重原作版权与维护公共利益之间达成高维度的私法秩序平衡。

Abstract: With the extensive application of generative artificial intelligence and deep learning algorithms in the field of cultural heritage preservation, AI restoration technology has become the core driving force for rescuing and revitalizing old motion picture film images. However, the deep intervention of AI technology has brought a fundamental shock to the traditional copyright system centered on “human authorship”. Grounded in the basic principles of China’s current Copyright Law, this paper first utilizes a technology categorization method to clarify the technical stratification of AI film restoration, dividing it into three levels: basic, moderate, and deep, which correspond to three legal characterizations: digital reproduction, derivative adaptation, and independent creation respectively. Furthermore, this paper deeply dissects the dogmatic dilemmas encountered in determining the “originality” of AI restoration outcomes, revealing the logical paradox between the aesthetic demand of “restoring the old as old” and the “novel expression” required by copyright law; it systematically analyzes the ownership disputes under the intertwined interests of algorithm providers, technical platforms, operators, and original copyright owners, as well as the risk of digital monopolies in the public domain. Finally, combining the legislative spirit of China’s Civil Code and Copyright Law, this paper proposes targeted institutional response paths: constructing a refined “subjective-objective dual-track” originality determination model; exploring the introduction of a “technical restoration right” or a neighboring rights protection mechanism specifically for digital film restoration outcomes; and improving the fair use restrictions and copyright collective management mechanisms for restored public domain images, thereby achieving a high-dimensional balance of private law order among encouraging technological investment, respecting original copyrights, and safeguarding public interests.