与单纯的技术失控和伦理失范相区分,生成式人工智能的创造性失控是指涌现出拟主体性的技术系统因其自主优化机制脱离人类预设伦理轨道而引发的系统性风险。生成式人工智能的创造性能力涉及认知科学、伦理学及存在论等多重维度,机器创造力不是对人类创造力的简单模仿或替代,而是技术系统与人类文明共同演化出的新型认知界面。创造性失控的本质是技术拟人化与伦理脱嵌化的冲突,表现为技术理性与价值理性的错位、技术社会化与责任归因的实践困境以及治理异步性与价值共识的制度悖论。这种矛盾冲突要求我们从技术重构、制度创新及价值共建三个维度来处理人机问题,通过嵌入式伦理治理框架、三元协同治理体系和敏捷治理生态的构建来重构技术拟人化伦理秩序。
The "Creative Out-of-control" in Generative Artificial Intelligence and the Reconstruction of Ethical Order In Technological Anthropomorphism
Wu Yingping(Southwest Minzu University)
Abstract Distinguished from mere technological runaway and ethical anomalies, the Creative Out-of- control in generative artificial intelligence (AI) refers to systemic risks arising when technological systems with quasi-subjectivity deviate from human-preset ethical trajectories through their autonomous optimization mechanisms. The creative capacities of generative AI span multiple dimensions encompassing cognitive science, ethics and ontology. Machine creativity does not merely imitate or replace human creativity, but represents a novel cognitive interface co-evolved through the symbiotic development of technological systems and human civilization. The essence of creative out-of-control lies in the conflict between technological anthropo‐ morphism and ethical disembeddedness, manifesting as the misalignment between technological rationality and value rationality, practical dilemmas in techno-socialization and accountability attribution, and institution‐ al paradoxes stemming from governance asynchronicity and value consensus. This paradigmatic confrontations require a tripartite approach to human-machine relations through technological reconfiguration, institutional innovation and value co-creation; it necessitates the establishment of an embedded ethical governance framework, a triadic collaborative governance system and an agile governance ecosystem to reconfigure the ethical order of technological anthropomorphization.
Key words artificial intelligence; Creative Out-of-control; technological anthropomorphism; ethical disembeddedness; order reconstruction; DeepSeek
■ 作者简介 吴映平,西南民族大学哲学学院副教授,四川 成都 610041。