【作者】 陈琳；

【导师】 宋艳；

【作者基本信息】 电子科技大学 ， 工商管理， 2022， 硕士

【副题名】基于亲社会价值观的调节

【摘要】 人工智能作为一项新兴的通用型技术,已经渗透到医疗、出行等各方面,正在深刻改变着我们的经济与社会。而与此同时,人工智能及其应用的负面影响和社会伦理问题也越来越突出,例如数据隐私、算法歧视、社会公平等。目前学术界对于人工智能伦理风险的研究,主要集中在成因、治理等宏观层面的探索,缺乏了解公众态度等微观层面上的实证研究。公众是技术产品消费的主体,亦是受科技负面效应直接影响的最大利益主体,其对人工智能技术产品风险的感知、评价和接受,可以最终决定人工智能技术的发展进程和未来方向。因此,公众对人工智能伦理风险的感知影响其参与风险治理的机制和边界条件是值得探讨的科学问题,具有一定的理论价值和实践意义。本研究将技术接受模型和风险感知理论作为理论支撑,以公众对于人工智能伦理风险的感知为自变量,公众参与风险治理的意愿为因变量,公众对科研机构、政府、企业、消费者保护协会和媒体的信任为中介变量,社会系统中科研机构、政府、企业、消费者保护协会和媒体的亲社会价值观导向为调节变量,构建研究模型,探讨四者之间的关系。通过三个阶段的问卷回收,以及数据整理,将1065份样本数据进行统计分析,实证检验了11个假设。结论为:(1)人工智能伦理风险感知对公众参与意向具有显著的负向影响;(2)公众对科研机构、政府和媒体的信任在人工智能伦理风险感知与公众参与意向之间起部分中介作用,对企业和消费者保护协会的信任的中介效应不显著;(3)社会系统中科研机构秉持的亲社会价值观导向在科研机构信任和公众参与意向之间起正向调节作用,政府和媒体秉持的亲社会价值观导向的调节作用不显著。基于以上的研究结论,本文分别从科研机构、政府、企业、非营利组织、媒体的主体层面提出了针对性的对策建议,以期为在人工智能技术领域构建“政府主导、专家引领、公众参与、社会监督”的治理新格局,以及企业负责任地创新提供科学依据。最后,考虑到目前研究的不足,本文提出了未来研究的方向。更多还原

【Abstract】 Artificial intelligence,as an emerging general-purpose technology,has penetrated into various aspects of healthcare and travel,and is profoundly changing our economy and society.However,at the same time,the negative impacts and ethical issues of AI and its applications are increasingly prominent,such as data privacy,algorithmic discrimination and social equity.Current academic research on the ethical risks of AI is mainly focused on macro-level explorations such as causes and governance,and lacks empirical studies at the micro-level such as understanding public attitudes.The public is the main consumer of technology products and the largest stakeholder directly affected by the negative effects of technology.Their perception,judgement and acceptance of the risks of AI technology products can ultimately determine the development process and future direction of AI technology.Therefore,exploring the mechanisms and boundary conditions by which the public’s perception of the ethical risks of AI affects their participation in risk governance is a scientific question of theoretical value and practical relevance.The thesis constructs a research model based on the technology acceptance model and risk perception theory.The relationship between the public’s perception of the ethical risks of artificial intelligence as the independent variable,the public’s willingness to participate in risk governance as the dependent variable,the public’s trust in research institutions,government,enterprises,consumer protection associations and the media as the mediating variable,and the pro-social value orientation of research institutions,government,enterprises,consumer protection associations and the media in the social system as the moderating variable are explored.Through three stages of questionnaire return,and data collation,the relevant hypotheses were finally tested empirically with a sample of 1065.The conclusions are:(1)perceived ethical risk of AI has a significant negative effect on public intention to participate;(2)public trust in research institutions,government and media partially mediates the relationship between perceived ethical risk of AI and public participation,and the mediating effect of trust in companies and consumer protection associations is not significant;(3)pro-social value orientation upheld by research institutions in the social system plays a positive role in the relationship between research institutions(3)The pro-social value orientation of research institutions in the social system acts as a positive moderator between trust in research institutions and public participation,whereas the pro-social value orientation of government and the media does not have a significant moderating effect.Based on the above findings,the thesis proposes targeted countermeasures at the levels of research institutions,government,enterprises,non-profit organisations and the media,with a view to providing a scientific basis for building a new governance pattern of "government-led,expert-led,public participation and social supervision",and for responsible innovation by enterprises.Finally,considering the shortcomings of the current study,the thesis proposes directions for future research.更多还原