恐怖谷理论
渲染(计算机图形)
通用人工智能
心理学
不可思议的
人机交互
计算机科学
简单(哲学)
认知心理学
人工智能
认知科学
社会心理学
认识论
精神分析
机器人
哲学
作者
Jan‐Philipp Stein,Markus Appel,Alexandra Jost,Peter Ohler
标识
DOI:10.1016/j.ijhcs.2020.102463
摘要
Digital technologies are advancing rapidly, growing to be more human-like and intelligent by the day. However, research shows that a machine's resemblance to humans can reach a critical level, which makes it seem uncanny to observers. While scholars have discussed this effect in terms of both human-like appearances and mental abilities, a potential interaction between the two aspects has hardly been addressed in literature. We designed a two-factorial experiment to overcome the identified research gap, introducing participants to digital agents with varying embodiment (text interface/human rendering) and mental capacity (simple algorithms/complex artificial intelligence). Our results show that the interaction of both factors indeed affects participants’ experience in a crucial way: Whereas an agent based on simple algorithms only evokes discomfort when embedded in a human-like body, the artificial intelligence is always perceived as eerie, regardless of its embodiment. Yet, additional findings raise doubts on the unidimensionality of participants’ affective response.
科研通智能强力驱动
Strongly Powered by AbleSci AI