摘要
在算法参与和替代人类驾驶活动的过程中,自动驾驶技术创新可能会与现有社会秩序和价值观产生冲突,引发包括行车安全、数据保护、算法歧视和就业替代在内的伦理风险。因对技术成熟度的不同理解和价值观方面的差异,不同社会群体之间可能出现伦理风险认知偏差,并表现出不同的技术包容度和接受度,进而通过影响产业政策取向延迟自动驾驶产业化进程。基于社会实验调查发现,我国社会公众对自动驾驶具有较高的宽容度和接受度,这为自动驾驶技术创新提供了良好的社会环境。通过实验组和对照组受访群体的对比分析发现,两类群体对不同类型自动驾驶伦理风险的认知存在显著差异,这表明是否拥有自动驾驶乘坐体验是影响社会公众伦理风险认知的关键因素。发展以人为本的人工智能,需要在加快技术创新的同时,通过更大范围的示范运营,消除公众对自动驾驶伦理风险的认知偏差,实现技术系统和社会系统协同演进,这是加快自动驾驶产业化发展的关键所在。
The integration of algorithms into driving systems,progressively replacing human control,may create friction with existing social norms and values.This technological shift potentially triggers ethical concerns including traffic safety,data protection,algorithmic discrimination,and workforce displacement.Divergent interpretations of technological maturity and conflicting value systems may foster cognitive biases in perception of the associated ethical risks across social groups.Such perceptual disparities manifest as divergent levels of public acceptance,potentially delaying autonomous vehicle commercializaiton by influencing industrial policy formulation.Our social experiment reveals that the Chinese public demonstrates notably high tolerance and acceptance rates toward AV technology,indicating a favorable sociotechnical environment for innovation.Crucially,significant percceptual gaps emerged between experimental and control groups in ethical risk perceptions across respondent categories,indicating that the experience of riding in autonomous rides fundamentally shapes ethical risk cognition.Advancing human-centered artificial intelligence requires dual-track progress:conducting large-scale demonstration operations and acclerating technological innovation.These initiatives must simultaneously drive technical advancement and address public cognitive biases regarding ethical risks.This co-evolution of technological and social systems emerges as a viable pathway for expediting autonomous vehicles industrilization.
作者
刘刚
郑凤阳
刘捷
LIU Gang;ZHENG Fengyang;LIU Jie
出处
《理论与现代化》
2025年第4期52-67,共16页
Theory and Modernization
基金
中国工程院战略研究与咨询项目“人工智能在社会科学研究领域的应用和问题研究”,项目编号:2024-HZ-25
教育部哲学社会科学实验室专项基金项目,项目编号:ZX20240380。
关键词
自动驾驶
伦理风险
认知偏差
示范运营
autonomous vehicles
ethical risks
cognitive biases
demonstration operation
