论人类和人工智能信任关系的三种研究路径

Three Research Paths on the Trust Relationship between Human and Artificial Intelligence

人工智能在人类社会发挥着愈加重要的作用,人工智能的专业能力使人类对其展现出信任态度。虽然人智之间不能按照人际的方式建立信任关系,但人工智能具有的自主能力表明它有可能回应人类的信任态度,因此基于人类中心主义视角将人工智能拒斥在信任关系之外的思路并不可取。“简化路径”虽然强调人智可以建立信任关系,但它涉及对信任构成要素的简化,因而会掩盖信任的本质特征。若将人智活动置于更广泛的交互背景中考察,人智之间可能建立间接信任关系并表现出对直接信任的“从属特征”,这为人智信任分析提供了可行路径。

Artificial intelligence(Al)is playing an increasingly important role in human society,its professional capabilities enable humans to express a trusting attitude towards it.Although it cannot establish trust relationships between human and Al in a interpersonal manner,the autonomy of Al indicates that it may respond to human trust attitudes.Therefore,the idea of excluding AI from trust relationships based on anthropocentrism is not advisable.The"simplified path"emphasizes that it can establish trust relationships between human and Al,but it involves simplifying the substantial elements of trust relationships,thus masking the essential characteristics of trust relationships.If human-Al interaction are examined in a broader context,indirect trust relationships between human and AI may be established,it will exhibit an"affliative feature"towards direct trust relationships,which provides a feasible path for thetrust analysis between human and Al.