生成式人工智能创造力的层级边界：符号重组、可能性搜索与意义激活

中文摘要

生成式人工智能的发展使“AI是否具有创造力”成为科技哲学与人工智能伦理的重要议题。既有观点或将AI生成视为机械拼接，或以产物新颖性承认其创造力，但均忽略创造力的层级差异。基于创造力理论、具身认知、人工智能哲学及意识层级视角，可将AI创造力区分为“符号重组—可能性搜索—意义激活”三层。AI在前两层具有显著优势，能够生成新形式、新方案和新假设；但因缺乏体感经验、情感意识、自我维持性欲求与责任承担能力，难以完成完整意义上的意义激活。由此，AI应被理解为具有生成性创造能力或可能性扩展能力，而不具备完整的主体性创造能力。AI时代人类创造力的价值将转向问题提出、意义赋予、价值筛选、人机协同与责任承担。

Abstract

The development of generative artificial intelligence has reopened the question of whether AI possesses creativity. Existing discussions often either deny AI creativity as mechanical recombination or affirm it solely on the basis of novel outputs. Both views overlook the hierarchical structure of creativity. From the perspectives of creativity theory, embodied cognition, philosophy of AI and layered consciousness, AI creativity can be divided into three levels: symbolic recombination, possibility search and meaning activation. Generative AI has strong capacities in the first two levels, generating new forms, solutions and hypotheses within large symbolic spaces. However, because it lacks embodied experience, affective consciousness, self-maintaining desire and responsibility-bearing capacity, it cannot complete meaning activation in the full human sense. AI should therefore be understood as possessing generative creativity or possibility-expanding capacity rather than full subject creative agency. The value of human creativity in the AI era lies in problem-posing, meaning-giving, value selection, human–AI collaboration and responsibility.

关键词：生成式人工智能；创造力；符号重组；意义激活；人工智能伦理

Keywords: Generative artificial intelligence; Creativity; Symbolic recombination; Meaning activation; Artificial intelligence ethics