人工智能监视技术伦理审查框架构建研究--基于文本分析

Construction of Ethical Review Framework of Artificial Intelligence Surveillance--Based on Text Analysis

聚焦人工智能监视技术的伦理监管困境,采用文本分析挖掘公众对AI监视的伦理关注,并进行整体和分场景比较,结合深度学习与神经网络对伦理属性进行情感分析。研究发现:公众关注的19个伦理问题贯穿AI监视技术的“动机-过程-结果”生命周期,关注程度在工作、教育和生活场景下存在差异;公众对AI监视整体持负面态度,对其动机合理性、过程正当性及监管有效性存在质疑;通过重要性-情感联合分析可以确定治理优先级,重要性等级高-情感值低的伦理问题需高度重视;最后,结合道义论与目的论,构建了AI监视的伦理审查框架。

Focusing on the ethical regulatory dilemma of AI surveillance,this paper uncovers the public's ethical concerns about AI surveillance by Text analysis,and compares them overall and by scenario.It combines deep learning and neural networks to perform sentiment analysis of the ethical factors.Results show that 19 ethical concerns of public span the"motivation-process-outcome"lifecycle of AI surveillance,and the concern level varies across work,education,and life scenarios.The public has a negative attitude towards AI surveillance in general,and questions the rationality of its motivation,the legitimacy of its process and the effectiveness of its regulation.The joint importance-sentiment analysis allows the prioritization of governance,with ethical issues of high importance and low sentiment requiring high priority.Finally,it constructs a framework for ethical review of Al surveillance by combining deontology and utilitarianism.