This chapter discusses AI potential in creative fields and how humans and generative AI creativity have recently been compared in the scientific literature. Generative AI models like ChatGPT have recently demonstrated remarkable capabilities in producing art, music, and literature. Research comparing human and AI creativity reveals that AI can match or surpass human performance in specific tasks. Some studies have found that the best humans can still outperform existing AI models regarding creative behaviours. Studies employing standardised creative tests have often shown that generative AI excels in tasks measuring fluency and originality, yet some aspects of creativity still see humans prevailing. The fact that AI excels and surpasses human performance in specific, experimentally controlled tasks does not automatically mean it will replace humans in creative jobs, which often require complex skills. Furthermore, AI's ability to produce creative products, such as artworks, does not diminish the subjective experience of human artists or the natural human tendency to communicate through art. However, AI-generated high-quality art that is fast and cheap to produce may pose economic challenges for human artists. Future research should focus on human-AI collaboration, exploring how AI can augment human creativity while addressing ethical considerations.

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

Creativity in the Age of AI: Comparing Human and Machine Performance Using Standardised Tests

  • Simone Grassini,
  • Selma Grødem,
  • Thea Gunnarskog,
  • Inga Tryggestad Sevre

摘要

This chapter discusses AI potential in creative fields and how humans and generative AI creativity have recently been compared in the scientific literature. Generative AI models like ChatGPT have recently demonstrated remarkable capabilities in producing art, music, and literature. Research comparing human and AI creativity reveals that AI can match or surpass human performance in specific tasks. Some studies have found that the best humans can still outperform existing AI models regarding creative behaviours. Studies employing standardised creative tests have often shown that generative AI excels in tasks measuring fluency and originality, yet some aspects of creativity still see humans prevailing. The fact that AI excels and surpasses human performance in specific, experimentally controlled tasks does not automatically mean it will replace humans in creative jobs, which often require complex skills. Furthermore, AI's ability to produce creative products, such as artworks, does not diminish the subjective experience of human artists or the natural human tendency to communicate through art. However, AI-generated high-quality art that is fast and cheap to produce may pose economic challenges for human artists. Future research should focus on human-AI collaboration, exploring how AI can augment human creativity while addressing ethical considerations.