Establishing trust between humans and intelligent technologies is crucial for a Human-AI Team (HAIT) in the workplace. Trust in a HAIT involves evaluating both the trustworthiness of team members and receiving feedback on one’s own trustworthiness. However, this evaluation and feedback process can be affected by inaccuracies due to subjective assumptions or errors leading to negative outcomes. The present study investigated how inaccurate trustworthiness feedback from human or AI team members affects individuals’ self-perceived trustworthiness and trust towards the feedback providor. An online scenario study with 511 participants showed that inaccurate trustworthiness feedback led to lower self-perceived trustworthiness compared to accurate trustworthiness feedback. However, no differences in trust in the team member giving the feedback were observed based on the source of the feedback (human vs. AI team member). These findings highlight the importance of inaccurate trustworthiness feedback in HAIT and suggest avenues for future research on trust dynamics in these new types of teams.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Feedback from AI Team Members: Implications on Self-image and Trust

  • Eleni Georganta,
  • Anna-Sophie Ulfert,
  • Zhen-Cong Ng

摘要

Establishing trust between humans and intelligent technologies is crucial for a Human-AI Team (HAIT) in the workplace. Trust in a HAIT involves evaluating both the trustworthiness of team members and receiving feedback on one’s own trustworthiness. However, this evaluation and feedback process can be affected by inaccuracies due to subjective assumptions or errors leading to negative outcomes. The present study investigated how inaccurate trustworthiness feedback from human or AI team members affects individuals’ self-perceived trustworthiness and trust towards the feedback providor. An online scenario study with 511 participants showed that inaccurate trustworthiness feedback led to lower self-perceived trustworthiness compared to accurate trustworthiness feedback. However, no differences in trust in the team member giving the feedback were observed based on the source of the feedback (human vs. AI team member). These findings highlight the importance of inaccurate trustworthiness feedback in HAIT and suggest avenues for future research on trust dynamics in these new types of teams.