互動語境如何影響大專院校生對機器人的感知、情緒與行為偏向

Abstract

Fiske 等（2007）及 Cuddy 等（2007）提出之（The stereotype content model）SCM 模型與（behaviors from intergroup affect and stereotypes）BIAS 模型指出，對他人溫暖與能力程度的感知會引起感知者不同的情緒與行為偏向（Fiske et al., 2002；Fiske et al., 1999；Cuddy et al., 2007），相關研究已將模型套用於人機互動，並確認其結果與人際互動相似（Mieczkowski et al., 2019；Reeves et al., 2020）。本研究相較於以往研究，加入工具型與互動型語境，探討不同的互動目的是否會影響人們對機器人的溫暖與能力的感知結果，以及其所引發的情緒和行為偏向是否與人際反應一致。 研究結果顯示工具型與互動型語境確實會影響人對機器人溫暖與能力的感知，尤其在工具型語境中，人對機器人的溫暖與能力感知程度更高。並且高溫暖高能力之機器人會引發人欣賞的情緒，並且引發主動與幫助的行為，低溫暖低能力之機器人則易引發人輕蔑的情緒，並且引發主動與被動的傷害行為。然而，嫉妒與同情等矛盾情緒在本研究中結果不明顯。 本研究結果顯示語境在討論人機互動時的重要性，並且證實人確實會對機器人產生溫暖與能力的感知，感知會進一步影響其情緒以及行為偏向。本研究結果有助於理解人感知機器人的機制，並為未來人機互動的研究與設計提供參考。

The SCM（The stereotype content model）and BIAS（behaviors from intergroup affect and stereotypes）model (Fiske et al., 2007; Cuddy et al., 2007) propose that perceptions of warmth and competence in others evoke distinct emotional and behavioral tendencies (Fiske et al., 2002; Cuddy et al., 2007). Prior studies have applied this to human-robot interaction, with findings similar to interpersonal results. (Mieczkowski et al., 2019; Reeves et al., 2020). This study adds instrumental and interactive contexts to examine how conversation goals affect warmth and competence perceptions of robots and whether these lead to responses akin to human interactions. Results show that context significantly influences perceptions: instrumental contexts enhance perceived warmth and competence in robots. Robots seen as high in warmth and competence elicit admiration and facilitation behaviors, while low-warmth, low-competence robots evoke contempt and harm behaviors. However, ambivalent emotions such as envy and pity were minimal. This study underscores the importance of context in human-robot interaction, demonstrating how perceptions of warmth and competence shape emotional and behavioral responses. These insights support future human-robot interaction research and design.