Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

In Scientific reports ; h5-index 158.0

Despite its considerable potential in the manufacturing industry, the application of artificial intelligence (AI) in the industry still faces the challenge of insufficient trust. Since AI is a black box with operations that ordinary users have difficulty understanding, users in organizations rely on institutional cues to make decisions about their trust in AI. Therefore, this study investigates trust in AI in the manufacturing industry from an institutional perspective. We identify three institutional dimensions from institutional theory and conceptualize them as management commitment (regulative dimension at the organizational level), authoritarian leadership (normative dimension at the group level), and trust in the AI promoter (cognitive dimension at the individual level). We hypothesize that all three institutional dimensions have positive effects on trust in AI. In addition, we propose hypotheses regarding the moderating effects of AI self-efficacy on these three institutional dimensions. A survey was conducted in a large petrochemical enterprise in eastern China just after the company had launched an AI-based diagnostics system for fault detection and isolation in process equipment service. The results indicate that management commitment, authoritarian leadership, and trust in the AI promoter are all positively related to trust in AI. Moreover, the effect of management commitment and trust in the AI promoter are strengthened when users have high AI self-efficacy. The findings of this study provide suggestions for academics and managers with respect to promoting users' trust in AI in the manufacturing industry.

Li Jia, Zhou Yiwen, Yao Junping, Liu Xuan

2021-Jun-30