<p>This paper addresses the question of whether artificial systems can be considered creative in the absence of cognition. It advances a novel perspective by grounding the analysis on a foundational premise: generative AI systems, such as LLMs and GMIs, are non-cognitive. This distinction is established through the application of the Minimal Cognitive Grid (MCG), offering a more precise entry point into the creativity debate. Despite their non-cognitive nature, these systems produce outputs that meet standard criteria for creativity—novelty and usefulness—and reproduce, in functional terms, the stages of human creative processes. A comparative analysis with the Wallas–Jaoui model supports this claim. However, the absence of intentionality and authenticity limits any attribution of genuine creativity. So, how can we define Artificial Creativity? To resolve this, the paper introduces a minimal definition of artificial creativity as a non-cognitive, non-intentional, and non-authentic generative mechanism. This is the first attempt to define the concept directly, rather than by exclusion. The definition clarifies the theoretical boundaries between natural and artificial creativity, avoids anthropocentric bias, and establishes a foundation for future research in computational creativity and philosophy of AI.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Artificial creativity: can there be creativity without cognition?

  • Matteo Da Pelo

摘要

This paper addresses the question of whether artificial systems can be considered creative in the absence of cognition. It advances a novel perspective by grounding the analysis on a foundational premise: generative AI systems, such as LLMs and GMIs, are non-cognitive. This distinction is established through the application of the Minimal Cognitive Grid (MCG), offering a more precise entry point into the creativity debate. Despite their non-cognitive nature, these systems produce outputs that meet standard criteria for creativity—novelty and usefulness—and reproduce, in functional terms, the stages of human creative processes. A comparative analysis with the Wallas–Jaoui model supports this claim. However, the absence of intentionality and authenticity limits any attribution of genuine creativity. So, how can we define Artificial Creativity? To resolve this, the paper introduces a minimal definition of artificial creativity as a non-cognitive, non-intentional, and non-authentic generative mechanism. This is the first attempt to define the concept directly, rather than by exclusion. The definition clarifies the theoretical boundaries between natural and artificial creativity, avoids anthropocentric bias, and establishes a foundation for future research in computational creativity and philosophy of AI.