[论文解读] An overview of research on human-centered design in the development of artificial general intelligence

Yang Yue, Joseph Z. Shyu|arXiv (Cornell University)|Aug 31, 2023

Artificial Intelligence in Healthcare and Education被引用 4

一句话总结

本文倡导在通用人工智能（AGI）开发中采用以人为本设计（HCD）方法，认为将尊严、隐私和自主性等人类价值观嵌入其中，对于实现伦理性和可持续性AGI至关重要。通过文献综述与内容分析，本文表明HCD可增强伦理对齐，并呼吁发展跨学科人才以支持负责任的AGI进步。

ABSTRACT

Abstract: This article offers a comprehensive analysis of Artificial General Intelligence (AGI) development through a humanistic lens. Utilizing a wide array of academic and industry resources, it dissects the technological and ethical complexities inherent in AGI's evolution. Specifically, the paper underlines the societal and individual implications of AGI and argues for its alignment with human values and interests. Purpose: The study aims to explore the role of human-centered design in AGI's development and governance. Design/Methodology/Approach: Employing content analysis and literature review, the research evaluates major themes and concepts in human-centered design within AGI development. It also scrutinizes relevant academic studies, theories, and best practices. Findings: Human-centered design is imperative for ethical and sustainable AGI, emphasizing human dignity, privacy, and autonomy. Incorporating values like empathy, ethics, and social responsibility can significantly influence AGI's ethical deployment. Talent development is also critical, warranting interdisciplinary initiatives. Research Limitations/Implications: There is a need for additional empirical studies focusing on ethics, social responsibility, and talent cultivation within AGI development. Practical Implications: Implementing human-centered values in AGI development enables ethical and sustainable utilization, thus promoting human dignity, privacy, and autonomy. Moreover, a concerted effort across industry, academia, and research sectors can secure a robust talent pool, essential for AGI's stable advancement. Originality/Value: This paper contributes original research to the field by highlighting the necessity of a human-centered approach in AGI development, and discusses its practical ramifications.

研究动机与目标

• 探讨以人为本设计在塑造伦理且可持续的AGI发展中的作用。
• 识别应指导AGI设计的关键价值观，如尊严、隐私和自主性。
• 填补AGI发展中关于伦理、社会责任及人才培育的实证研究空白。
• 推动产业、学术界与研究领域之间的跨学科合作，以促进AGI治理。

提出的方法

• 对AGI与以人为本设计相关的学术与产业资料进行全面文献综述。
• 通过内容分析提取与AGI开发中HCD相关的主要主题与概念。
• 评估AGI研究中现有的理论、最佳实践与伦理框架。
• 将价值整合（如共情与社会责任）的发现综合进AGI系统设计中。
• 通过分析现有研究与产业趋势，识别AGI人才发展中的关键挑战。
• 提出一种以人类价值观和跨学科合作为优先的治理模型。

实验结果

研究问题

• RQ1 如何有效将以人为本设计原则整合到通用人工智能的开发中？
• RQ2 哪些核心人类价值观——如尊严、隐私和自主性——应指导AGI系统的伦理设计？
• RQ3 当前关于AGI发展中伦理、社会责任及人才培育的研究存在哪些关键空白？
• RQ4 产业、学术界与研究机构之间的跨学科合作如何促进AGI的伦理进步？

主要发现

• 以人为本设计对于确保AGI系统与人类价值观一致至关重要，尤其是尊严、隐私与自主性。
• 在AGI开发中融入共情、伦理与社会责任，可显著提升此类系统的伦理化部署。
• 亟需开展更多聚焦于AGI语境下伦理、社会责任及人才培育的实证研究。
• 跨学科举措对于培养支持AGI可持续且负责任发展的必要人才至关重要。
• 产业、学术界与研究领域的协作对于建立坚实的AGI人才储备池必不可少。
• 将以人为本的价值观融入AGI设计，可促进长期的伦理与可持续技术进步。