Human-centered AI (HCAI) concerns itself with prioritizing human needs and values predominantly in work-oriented settings, safety-critical situations, and in terms of societal impact. For example, the responsible use of AI in healthcare decision-making or the ethical use of AI bots in social media. The Arts provide an alternative lens through which to examine the challenges of human-centered AI. In this chapter, HCAI Grand Challenges are revisited from an Arts perspective. Generative AI (GenAI) is frequently used in the Arts, raising HCAI questions about how to design and evaluate GenAI systems that are human-centered focusing on human agency, aesthetic control, and inspiration. At the same time, questions are raised about how these systems can be trained ethically and responsibly to reduce bias. Questions are also raised about how HCAI techniques can be developed to evaluate AI systems in the Arts where uncertainty and surprise might be more important than transparency of explanation. Examples from research and artistic practice are used throughout this chapter to exemplify and bring to life current discourses and approaches to HCAI and the Arts.

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

Human-Centered AI in Arts

  • Nick Bryan-Kinns,
  • Phoenix Perry,
  • Elizabeth Wilson,
  • Anna Wszeborowska

摘要

Human-centered AI (HCAI) concerns itself with prioritizing human needs and values predominantly in work-oriented settings, safety-critical situations, and in terms of societal impact. For example, the responsible use of AI in healthcare decision-making or the ethical use of AI bots in social media. The Arts provide an alternative lens through which to examine the challenges of human-centered AI. In this chapter, HCAI Grand Challenges are revisited from an Arts perspective. Generative AI (GenAI) is frequently used in the Arts, raising HCAI questions about how to design and evaluate GenAI systems that are human-centered focusing on human agency, aesthetic control, and inspiration. At the same time, questions are raised about how these systems can be trained ethically and responsibly to reduce bias. Questions are also raised about how HCAI techniques can be developed to evaluate AI systems in the Arts where uncertainty and surprise might be more important than transparency of explanation. Examples from research and artistic practice are used throughout this chapter to exemplify and bring to life current discourses and approaches to HCAI and the Arts.