Scholarlink

<p>This paper examines the ontological status of images generated through AI-powered platforms such as Midjourney and DALL-E, termed “generative art.” We argue that artworks are functional kinds. In the production of art, there is usually a <i>creator</i> who uses a <i>medium</i> to create a <i>work</i> that is available for <i>appreciation—</i>a sequence we call CMWA. We argue that artworks are artifacts made for appreciation. Generative art often fits this functional characterization, so it is functionally continuous with traditional art at a macro level. At a micro level, there are significant functional differences. With generative art, the boundary between creator, medium, and work are harder to discern because the medium (a computer algorithm) actively contributes to the outcome. This can be regarded as a kind of collaboration, but it differs from traditional art, since the human artist shapes generative art through prompting. We call this “catalytic collaboration”. In addition, rather than simply reflecting the skills, context, and memory of an individual human creator, generative art is informed by the stored records of everything in its vast training corpus. We call this “hypermnesia”. Thus, when we appreciate a work of generative art, we are not focused exclusively on the creative individual. Each work also reflects large swaths of cultural memory. The functional analysis given by the CMWA sequence is, thus, both continuous with and different from traditional art.</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

From canvas to code: lessons from generative art

  • 期刊论文

  • 发表日期：

    2026年2月4日

查看全文
  • Maedeh Norouzi,
  • Jesse Prinz

摘要

This paper examines the ontological status of images generated through AI-powered platforms such as Midjourney and DALL-E, termed “generative art.” We argue that artworks are functional kinds. In the production of art, there is usually a creator who uses a medium to create a work that is available for appreciation—a sequence we call CMWA. We argue that artworks are artifacts made for appreciation. Generative art often fits this functional characterization, so it is functionally continuous with traditional art at a macro level. At a micro level, there are significant functional differences. With generative art, the boundary between creator, medium, and work are harder to discern because the medium (a computer algorithm) actively contributes to the outcome. This can be regarded as a kind of collaboration, but it differs from traditional art, since the human artist shapes generative art through prompting. We call this “catalytic collaboration”. In addition, rather than simply reflecting the skills, context, and memory of an individual human creator, generative art is informed by the stored records of everything in its vast training corpus. We call this “hypermnesia”. Thus, when we appreciate a work of generative art, we are not focused exclusively on the creative individual. Each work also reflects large swaths of cultural memory. The functional analysis given by the CMWA sequence is, thus, both continuous with and different from traditional art.