公众人工智能价值敏感性的测度解析
Analysis on Measurement of Value Sensitivity with Artificial Intelligence Among Public
张学义 1周洲 2郑宇涵2
作者信息
- 1. 东南大学哲学与科学系,江苏南京,211189;江苏省道德发展智库,江苏南京,211189
- 2. 东南大学哲学与科学系,江苏南京,211189
- 折叠
摘要
"价值敏感性"源于巴迪亚·弗里德曼的"价值敏感性设计"概念,指在技术设计、研发、生产和使用中具有伦理意蕴的价值感知能力.课题组创新性地应用"三方方法论",通过多指标多因素结构方程模型(MIMIC)测量公众对人工智能的价值敏感性,评估现有技术是否支持价值敏感设计目标.研究发现:(1)公众对人工智能的价值敏感排序为:透明性、算法公正、劳动就业保障、责任归属等,说明这些价值需在技术设计阶段予以重视;(2)信息安全和生态失衡未引起广泛关注;(3)性别和职业对价值敏感性无显著影响,但年龄和学历差异显著.
Abstract
"Value sensitivity"originates from Batya Friedman's concept of"Value-Sensitive Design",referring to the ability to perceive ethically relevant values during the design,development,production,and use of technology.The research team innovatively applied the"tripartite methodology"and used the Multiple Indicators and Multiple Causes(MIMIC)structural equation model to measure public value sensitivity toward artificial intelligence(AI)and assess whether the current technologies support value-sensitive design goals.The study found that:(1)Public sensitivity to AI values is ranked as transparency,algorithmic fairness,job security,and responsibility attribution,indicating that these values should be prioritized in the design stage;(2)Information security and ecological imbalance received less attention;(3)Gender and profession showed no significant impact on value sensitivity,but age and education level did show significant differences.
关键词
人工智能/价值敏感性/多指标多因素模型/伦理治理Key words
Artificial intelligence/Value sensitivity/MIMIC/Ethical governance引用本文复制引用
出版年
2025
NETL
NSTL
万方数据