Rethinking the Mindfulness and Mindlessness in Media Evocation of Human-Machine Communication: A Case Study on the In-Car Robot "Nomi"

Source: Bing Wang, Longxiang Luo, Weizi Liu; Human-Machine Communication 12, journal article (open access) · 2026

The commonly used Media Equation framework in human-machine communication research exhibits limitations when explaining human-machine relationships in the AI era, giving rise to the Media Evocation paradigm. Based on the Media Evocation framework, this study employs in-depth interviews to investigate whether interactions between humans and the in-car robot Nomi are mindful or mindless, while also examining whether these interactions prompt reflections on the machine’s ontology. The findings indicate that most participants perceive Nomi as a human-like actor and interact with it mindfully. Highquality social cues, such as voice and facial expressions, can trigger mindful anthropomorphic responses in human-machine interaction, eliciting behaviors such as decorating the robot and projecting emotions onto it. Most participants regard Nomi as occupying a hybrid social role, akin to a combination of a pet and a friend, and perceive it as an independent entity rather than one that is reliant on humans. Mindfulness and Mindlessness exhibit variations in distinguishing criteria, triggering conditions, and fundamental characteristics. Discussions of the ontology of machines need to be situated within specific relationships and contexts, reflecting dynamic, complex, and variable characteristics.

Introduction. Human-Machine Communication (HMC) research typically relies on the Media Equation as a foundational theory (Fortunati & Edwards, 2022; van der Goot, 2022), emphasizing that users unconsciously treat computers and new media like humans by adopting social attitudes, behaviors, and rules (Reeves & Nass, 1996). A series of subsequent experiments were later summarized under the notion that “computers are/as social actors,”1 and this framework attributes such interactions to mindlessness, which are automatic responses to technological social cues without deliberate reflection on the machine’s nature (Kim & Sundar, 2012; Lombard & Xu, 2021). However, the proliferation of advanced AI technologies has transformed humanmachine relationships. Intelligent machines like social robots and voice assistants now possess highly anthropomorphic features, transcending the functional limitations of early computers (van der Goot & Etzrodt, 2023; B. Wang et al., 2024). This evolution exposes critical limitations of the Media Equation: it overlooks user introspection, avoids ontological questions about machines (Etzrodt & Engesser, 2021; Fortunati & Edwards, 2021), and fails to account for mindful engagement with AI systems. To address these gaps, the Media Evocation paradigm has been proposed among scholars of HMC, conceptualizing intelligent machines as philosophical provocateurs that prompt reflection on human-machine ontological boundaries and frame interactions as mindful (van der Goot & Etzrodt, 2023). The validity of this paradigm in explaining complex AI-era HMC remains underexplored. Nomi, the voice-based assistant of the Chinese new energy vehicle brand NIO, is touted as the world’s first in-vehicle AI. Nomi features a humanoid face-like interface with oval eyes that swivel and blink to address each occupant directly. Nomi can accomplish tasks like information inquiry and in-vehicle device control through voice-based communication and engage in socioemotional interactions with the driver and passengers. Drivers can also customize Nomi by giving it a new name, and more importantly, they dress Nomi in various clothes and hats, treating it as a social actor. The act of dressing up machine products has sparked social reflection. It seems that people perceive Nomi as not only like but actually being a real social member, and they appear to be deeply engaged in enhancing the human-machine interaction experience mindfully, aligning with the assumptions of the Media Evocation. Clarifying whether such interactions are mindful or mindless, and whether they prompt ontological reflection, has important implications for HMC theory and AI design. This study draws on in-depth interviews with 17 Nomi users to address two central objectives: (1) to examine the nature (mindful vs. mindless) and manifestations of human- Nomi interaction; and (2) to assess whether these interactions elicit reflections on Nomi’s ontological status. By exploring both dimensions, the study aims to unpack the complexity of human-Nomi interactions and evaluate the extent to which the Media Evocation paradigm can account for emerging phenomena in HMC.

Related work. The Media Equation emerged in the 1990s, supported by a series of experiments that people naturally and socially respond to media. The public is aware that such responses are irrational and often deny that they would ever react this way, refusing to acknowledge that they treat media, such as computers, as having life or equal social status to humans. The most well-known research in the Media Equation series is the CASA paradigm, which posits that people perceive media like computers as social actors and interact with them according to social rules (Nass et al., 1994). Rooted in Media Equation, HMC researchers have long discovered a series of cases where individuals exhibit social behaviors toward computers and other new media, involving factors such as similarity (Moon & Nass, 1996), gender stereotypes (Nass et al., 1997), reciprocity (Fogg & Nass, 1997), and politeness (Nass et al., 1999). As more intelligent products equipped with the latest AI technology enter people’s lives, the Media Equation theory has shown certain limitations in explaining these emerging human-machine interactions. More specifically, first, this theory has paid relatively little attention to users’ introspection and reflection (Reeves & Nass, 1996). Second, it does not delve into the ontological issues concerning machines, in other words, ontological questions such as What is a machine?, and How do the abilities of humans and robots as communicators differ? remain unresolved (Fortunati & Edwards, 2021), which left difficulties for people to determine whether a machine is a thing/object or person/subject (Etzrodt & Engesser, 2021). Third, except for the mindlessness, our understanding remains insufficient regarding the extent to which and the mechanisms through which novel conversational AIs and embodied systems elicit mindful forms of engagement that challenge, contest, or reconfigure the categorical boundaries traditionally drawn between humans and machines. In comparison to the Media Equation, Media Evocation focuses on the mindful responses of individuals to machines, centering on how people rethink themselves and machines through communication with machines, and it emphasizes reflections and negotiations regarding the ontological boundaries concerning humans and machines (van der Goot & Etzrodt, 2023). This paradigm is influenced by the concept of evocative objects (Turkle, 2005, 2011). An evocative object is considered an object situated between any diametrical opposites, such as humans and other objects. It can evoke reflection and discussion on the essence of related subjects, referring more specifically to one’s own and human identity (van der Goot & Etzrodt, 2023).

Method. Much HMC research, which is influenced by the Media Equation, emphasizes instrumental processing and cue-response dynamics rather than symbolic or relational interaction (Edwards & Etzrodt, 2025), and employs quantitative research methods such as experiments and surveys. However, with the emergence of Media Evocation and the growing attention to meaning and ontology, two key insights have emerged: First, meaning does not emerge automatically; instead, it is formed through symbolic projection and contextual negotiation (Edwards & Etzrodt, 2025); second, scholars are advocating for the adoption of more qualitative research methods (Guzman, 2023; Spence et al., 2023; van der Goot & Etzrodt, 2023). In this study, we adopt in-depth interviews, aiming to capture specific scenarios of human-machine interaction through conversations with car owners. These interviews allow for greater attention to the relationship between humans and Nomi as described by the owners, and also enable exploration of participants’ ideas regarding more abstract concepts, such as the ontology of machines. During the development of the interview protocol, the content was systematically designed based on the research objectives and research questions. The interview outline encompassed five major areas: (1) participants’ demographic information, (2) their everyday experience and use of Nomi, (3) assessments of mindfulness and mindlessness in human-Nomi interactions, (4) participants’ perceptions and intentional behaviors related to the anthropomorphism of Nomi, and (5) ontological reflections on the Machine. In addition, supplementary questions were incorporated as interviews progressed to allow deeper exploration of emerging topics and meanings. The full interview protocol is provided in Appendix A. To examine mindfulness, mindlessness, and anthropomorphism in participants’ interactions, this study draws on prior research by van der Goot (2022), Kim and Sundar (2012), and Powers and Kiesler (2006). In qualitative research, findings are typically presented through rich description, which foregrounds the narrative dimension of inquiry, relies on careful contextualization of social behavior, and emphasizes nuance, detail, and the underlying structures of meaning. Rooted in this epistemological orientation, Media Evocation emphasizes a relational logic, where meaning does not emerge automatically but is instead constructed through symbolic projection and contextual negotiation. Building upon these theoretical and methodological premises, the present study analyzes car owners’ narratives to assess whether their interactions with Nomi reflect mindful or mindless engagement, and to interpret their anthropomorphic behaviors. The operationalization of these concepts is detailed in the table on the following page. Based on the related research, mindful interactions are generally characterized as reflective, intentional, and involving agentive attribution, whereas mindless interactions refer to social responses triggered by cues despite individuals’ explicit denial of the machine’s social status. In this study, the analysis of mindfulness focuses on whether participants consciously acknowledge Nomi as a socially agentive, life-like entity with its own personality, a judgment that requires deliberate reflection. By contrast, Mindlessness is examined through indirect questioning about participants’ social behaviors toward Nomi, and instances in which such behaviors co-occur with denials of Nomi’s social role are interpreted as mindless and unintentional responses. At a deeper level, questions related to anthropomorphism and ontology take a more direct approach, probing how participants reason about Nomi’s status during actual interactions.

Using thematic analysis, a table-based method that considers both theoretical and operational aspects, researchers can systematically organize and describe a data set in detail and clarify multiple dimensions of the research topic while emphasizing recurring patterns of meaning (Braun & Clarke, 2006; B. Wang et al., 2024).

Discussion. Rethinking Mindfulness and Mindlessness: The Prospect of a New Cognitive Trajectory The core of this study lies in the exploration of Mindfulness and Mindlessness. An analysis of the research findings reveals the coexistence of both mindful and mindless interactions of humans and machines. Specifically, while a small portion of respondents still exhibit mindless behavior, the majority demonstrate characteristics of mindful behavior. High-quality social cues play a significant role in this shift. As user experience improves, there is a growing understanding of products like robots, highlighting the interplay between mindful and mindless mechanisms (Lombard & Xu, 2021). Contrary to previous research findings, this study reveals that people are increasingly inclined toward mindful behavior. From the perspective of human-machine communication, the boundary between mindfulness and mindlessness has become less clear. The following discussion elaborates on this based on the study findings. First, within the HMC framework, the criteria for distinguishing mindfulness from mindlessness have become blurred. According to the assumptions of the Media Equation, the standard for mindlessness is that users respond socially based on cues provided by the machine, thereby following rules of human-like interaction while denying the machine’s status as a social actor (Kim & Sundar, 2012; Nass & Moon, 2000). However, this study reveals that individuals do not engage with Nomi mindlessly, as they view Nomi as possessing its own life and acting socially akin to a human. Driven by both emotional responses and rational reflection, most respondents do not view Nomi as merely a machine, whereas only a small group denies its social status while still holding social expectations for it. Rooted in emotional dependence and attachment accumulated over prolonged interaction, this perspective highlights the need to examine human–machine relationships in naturalistic, long-term contexts. Consequently, what previous studies emphasized regarding mindfulness, which is recognizing machines as social actors, has now become mainstream. Previous studies have suggested that deliberation serves as a criterion for distinguishing mindfulness and mindlessness (Złotowski et al., 2018). The distinction lies in whether individuals respond socially subconsciously by following ingrained thought patterns or engage in reflective consideration before exhibiting corresponding behaviors. Although individuals today may still not consciously deliberate before interacting with Nomi, their natural responses are shaped by accumulated experiences from the era of human-machine interaction. These responses represent social behavior that acknowledges the machine’s agency, differing from the traditional criteria used in HMC research to judge mindfulness and mindlessness. Essentially, mindfulness and mindlessness relate to human understanding of social interaction. In this study, the concept of human understanding in humanmachine relationships needs to be reconsidered based on the HMC-informed criteria for distinguishing mindfulness and mindlessness. Second, considering anthropomorphic strategies and social cues as one of the triggers of mindfulness and mindlessness, differences from previous cues also emerge. Enhancements to the visual and auditory aspects of intelligent robots facilitate harmonious human-machine relationships. Facial expressions and gaze are common nonverbal behaviors for establishing rapport (Seo et al., 2018). For Nomi, the feedback mechanism represents its most significant advantage over other products, as gaze and appropriate facial expressions make car owners perceive communication as bidirectional.

Conclusion. This study is grounded in the Media Evocation paradigm and engages in an exploration of mindfulness and mindlessness, empirically validating the framework proposed by van der Goot and Etzrodt (2023). Unlike previous research (Nass & Moon, 2000), it finds that people exhibit more mindful social responses toward machines. Nass and Moon once noted that “people can be trained to be more mindful of context cues” (Nass & Moon, 2000, p. 98) and may break free from primitive, automatic responses through deliberation. However, they argued that such a strategy tends to complicate the process and is neither typical nor common. In contrast, the present study finds that in the AI era, people have begun to consciously recognize the social nature of Nomi, making it a widely shared understanding. In the meantime, Mindfulness and Mindlessness show changes in terms of distinguishing criteria, triggering conditions, and their fundamental nature, with high-quality social cues often eliciting more proactive anthropomorphic behaviors. These findings suggest the need to reconceptualize mindfulness and mindlessness in human–machine interaction in the era of AI. Whereas previous studies primarily focused on how interpersonal social cues and stereotypes shape human–machine relationships, the current era witnesses a new understanding of human engagement with machines. Breaking away from anthropocentric perspectives and developing a new mode of human–machine thinking may be essential to advancing the discussion on this topic. The ontological discussion of machines needs to be situated within relationships, viewed through a dynamic lens, and grounded in specific contexts (Edwards & Etzrodt, 2025; Suchman & Thimm, 2024). Most individuals now recognize machines embedded with advanced AI technologies as social agents capable of forming emotional relationships with humans, and their ontological understanding of machines has shifted through frequent human-machine interactions. When responding to interview questions, participants often associated the role played by Nomi with specific contexts while simultaneously using ambiguous or indeterminate terms to describe its ontology.

Limitations. Although scholars of Media Evocation emphasize the need for qualitative methods in HMC research (van der Goot & Etzrodt, 2023), experimental methods can also play an important role, particularly in measuring variables related to abstract concepts such as Mindfulness and Mindlessness. In this study, these abstract concepts were operationalized concretely through interviews, but the results may be influenced by the relationship between interviewer and interviewee. Future research could consider the use of mixed methods.

In addition, this study does not fully account for the influence of individual and contextual factors on people’s responses (Lombard & Xu, 2021). Different users may exhibit mindful or mindless reactions toward Nomi due to various individual factors—such as gender, personality traits, prior usage experience, and level of knowledge, or contextual factors, including the specific usage setting or family dynamic. These distinctions were not examined in the present study, which constitutes an important direction for future research.