Page 162 - ISC PROCEEDINGS 21.4

anthropomorphism, expressive language, and appearance shape consumer responses
(Gao et al., 2023; Zhou et al., 2024). In contrast, consumer-centered studies emphasize
the role of psychological factors and individual characteristics, including perceived ease of
use, empathy, and trust toward virtual streamers. Overall, regardless of the analytical
perspective employed, most previous studies suggest that virtual streamer-related factors
tend to elicit positive consumer responses, such as increased acceptance, impulse
purchasing intentions, and online engagement.
However, within the technology research domain, numerous studies have
demonstrated that negative user responses also play a critical role in explaining
behavioral outcomes and the formation of perceptions toward emerging technologies (Ali
et al., 2016; Lapointe & Rivard, 2005). In the context of livestream commerce, virtual
streamers are no exception. Although some recent studies have begun to explore
negative reactions, such as aversion toward virtual streamers (Xiao et al., 2025) or
resistance to AI participation in livestream commerce (Shao & Ho, 2025), this line of
research remains relatively limited. Moreover, most existing studies have been conducted
in countries with highly developed livestream commerce ecosystems, most notably China,
while the context of developing countries such as Vietnam remains largely underexplored.
Therefore, examining consumers’ negative responses, particularly resistance intentions
toward virtual streamers, is necessary to provide a more comprehensive understanding of
the impact of this technology in livestream commerce.
In this study, user resistance is not considered a random reaction but rather a
consequence of how users perceive the core attributes of AI agents and the potential
risks they introduce. Two key characteristics of virtual streamers that shape interaction
quality are perceived autonomy and perceived anthropomorphism. Although these
attributes may enhance the shopping experience by creating a more vivid and realistic
interaction environment, they may also trigger psychological discomfort. When
interacting with digital entities that exhibit high autonomy and human-like characteristics,
users often conduct cognitive evaluations of potential losses. In this regard, the present
study focuses on two prominent forms of perceived risk include intrusiveness risk and
privacy disclosure risk to better explain the psychological mechanisms through which
users develop resistance intentions toward virtual streamers in livestream commerce.
2. Theoretical background, literature review and methodology
2.1. Theoretical background and literature review
This study is grounded in the Uncanny Valley Theory and Psychological Reactance
Theory to explain users’ emotional and behavioral responses when interacting with AI
agents. The Uncanny Valley Theory, proposed by Mori (1970), suggests that people’s
positive feelings toward artificial entities increase as those entities become more human-
like. However, when an artificial agent appears almost human but still contains subtle
unnatural characteristics, users may experience feelings of eeriness, discomfort, or
distrust. This reaction often occurs when there is an inconsistency in the level of human
likeness, such as asynchronous facial expressions or unnatural movements (MacDorman
& Chattopadhyay, 2016; Tinwell et al., 2013). In addition, the perceived autonomy of
artificial agents may intensify the uncanny valley effect when users believe that the agent
is capable of acting independently (Stein & Ohler, 2017). According to the Psychological
Reactance Theory developed by Brehm (1966), individuals tend to react negatively when
they perceive that their freedom of choice is threatened. When individuals believe that
their autonomy is being restricted, they may experience psychological reactance, which

161