Perceived and Measured Task Effectiveness in Human-AI Collaboration
CHI '20: CHI Conference on Human Factors in Computing Systems Honolulu HI USA April, 2020(2020)
摘要
Human-AI Collaboration is emerging all around with the increasing utilisation of AI. Few prior studies have investigated the perceived effectiveness of users solving tasks with AI. To expand on these, we conducted a within-subjects repeated measures study involving 35 participants sorting household waste according to recyclability both with and without the help of an AI system. Our results show that people both sorted more effectively and perceived themselves more effective. Furthermore, we document a trend where people sorting without suggestions perceived themselves more effective than they were, while the opposite was true for people when sorting receiving suggestions. Based on our results we propose open questions for future research on perceived effectiveness when collaborating with AI systems.
更多查看译文
关键词
Human-Al collaboration, Artificial Intelligence, Lab study, Effectiveness
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络