Computers Are Social Actors Research Articles

Transitioning to Human‐Centered AI: A Systematic Review of Theories, Scenarios, and Hypotheses in Human‐AI Interactions

ABSTRACTThis study conducted a systematic review of human‐AI interaction (HAI)over the past decade for the implemented theories and scenarios, and the tested hypotheses to discover the changes in the current transition to human‐centered AI (HCAI). Moving from acceptance theories, Computers are social actors (CASA), anthropomorphism, and the integrative trust model are the most frequent theories. Augmentation scenarios of decision‐making, teamwork, and human‐AI collaborations are common in the latest HAI studies. Users' trust, acceptance, and intention to use an AI system are the main research targets in HAI studies. These trends show a clear transition toward HCAI. This paper also discusses opportunities tied to HAI studies based on the interconnections between the various theories, scenarios, and hypotheses.

What, Was She Not Human? The Mediated Effects of Virtual Influencers’ Identity Disclosure Timing on Behavioral Intentions: Focusing on the Moderating Role of Influencer-Brand Fit

In the current era, numerous human-like virtual influencers (VI) on social media promote brands without immediately revealing their non-human nature, intending for viewers to realize it later. This marketing strategy raises a question regarding their endorsement effectiveness. Grounded in Computers are Social Actors (CASA) paradigm, this online experiment employs a quota sample of Korean adults (N = 128) to explore how the timing of VI identity disclosure (before vs. after exposure to sponsored content) affects viewers’ attitudes toward the VI and, subsequently, their intentions to purchase the endorsed brand or spread word-of-mouth. Additionally, based on the match-up hypothesis, this study examines how perceived influencer-brand fit impacts these relationships. Moderated mediation analyses reveal that viewer attitudes mediate the relationship between disclosure timing and behavioral intentions, with VI-brand fit significantly moderating this mediation. The findings underscore the negative effects of delayed VI identity disclosure and emphasize the critical importance of a strong influencer-brand match.

Should government chatbots behave like civil servants? The effect of chatbot identity characteristics on citizen experience

Government chatbots are naturally imparted the identity of civil service agents, because they replace human civil servants in providing services on the frontline. However, little is known about whether government chatbots should present key characteristics consistent with their identity. Based on Computers Are Social Actors (CASA) theory and the stereotype content model (SCM), this study explored the effect of various identity characteristics on citizens' experiences in Chinese government chatbots. Valid data were obtained from 705 respondents and analysed using structural equation model, we found that manners, proactivity, conscientiousness, fairness, and professionalisation significantly influence the citizens' experiences, and that task-oriented language style has a negative effect. These effects are partly mediated by perceptions of warmth and competence. These findings contribute theoretically to the literature on chatbot identity and provide implications for incorporating chatbots into the design of e-government context.

“Sorry, I Don’t Understand …”: effects of task type, personality presentation and performance error on user trust of a personal digital assistant

PurposeBased on the theoretical predictions of media equation theory and the computers-are-social-actors (CASA) perspective, this study aims to examine the effects of performance error type (i.e. logical, semantic or syntactic), task type and personality presentation (i.e. dominant/submissive and/or friendly/unfriendly) on users’ level of trust in their personal digital assistant (PDA), Siri.Design/methodology/approachAn experimental study of human–PDA interactions was performed with two types of tasks (social vs functional) randomly assigned to participants (N = 163). While interacting with Siri in 15 task inquiries, the participants recorded Siri’s answers for each inquiry and self-rated their trust in the PDA. The answers were coded and rated by the researchers for personality presentation and error type.FindingsLogical errors were the most detrimental to user trust. Users’ trust of Siri was significantly higher after functional tasks compared to social tasks when the effects of general usage (e.g. proficiency, length and frequency of usage) were controlled for. The perception of a friendly personality from Siri had an opposite effect on social and functional tasks in the perceived reliability dimension of trust and increased intensity of the presented personality reduced perceived reliability in functional tasks.Originality/valueThe research findings contradict predictions from media equation theory and the CASA perspective while contributing to a theoretical refinement of machine errors and their impact on user trust.

My avatar makes me feel good? The effect of avatar personalisation and virtual agent interactions on self-esteem

ABSTRACT The theory of Objective Self-Awareness (OSA) and related studies suggest that embodiment of personalised avatars may induce self-awareness and influence self-esteem. Additionally, the Computers Are Social Actors (CASA) paradigm suggests that humans may mindlessly respond to computers in ways that are similar to human interactions. Based on those assertions, it is plausible that virtual embodiment of a personalised avatar and interactions with a virtual agent can shift self-esteem. However, those effects on self-esteem have not been thoroughly examined in past studies. To address these research gaps, a 2 (avatar personalisation: personalised vs. non-personalised avatar) × 2 (virtual agent interaction valence: positive vs. negative) between-subjects experiment was conducted using a Virtual Reality (VR) simulation (N =171). Findings from the study showed that there was no effects of avatar personalisation and virtual agent interaction valence on state self-esteem change. However, the pairwise comparisons present some preliminary indications that avatar personalisation and positive interactions with a virtual agent may facilitate improvements in state self-esteem altogether. Implications of the study findings are discussed.

Beyond learning with cold machine: interpersonal communication skills as anthropomorphic cue of AI instructor

Prior research has explored the impact of diverse anthropomorphic interventions on the effectiveness of AI (artificial intelligence) instructors. However, the exploration of interpersonal communication skills (e.g., self-disclosure) as anthropomorphic conversational cues for AI instructors is rare. Considering the positive impact of the self-disclosure of human instructors and guided by the social penetration theory (Altman & Taylor, 1973) and computers are social actors (CASA) paradigm (Nass & Moon, 2000), this study explores the role of self-disclosure by AI instructors and the mediating role of emotional attachment between AI instructors’ self-disclosure and students’ learning experiences (learning interest and knowledge gain). Additionally, it examines the differences in students’ emotional attachment, learning interest, and knowledge gain between AI and human instructors. Through a 2 (AI instructor vs. human instructor) × 2 (self-disclosure: yes or no) experiment, this study concluded that 1) consistent with human instructors, self-disclosure by AI instructors led to higher emotional attachment, learning interest, and knowledge gain; 2) emotional attachment played an important mediating role in AI instructor self-disclosure and students’ learning interest and knowledge gain; and 3) in the context of self-disclosure, students exhibited similar levels of emotional attachment to both AI and human instructors, with no significant differences observed. Regarding learning outcomes, while students demonstrated a greater interest in learning during courses taught by AI instructors, the difference in knowledge gained from AI and human instructors was not significant. The results of this study contribute to the understanding of the anthropomorphic cues of AI instructors and provide recommendations and insights for the future use of AI instructors in educational settings.

Emojifying chatbot interactions: An exploration of emoji utilization in human-chatbot communications

The prevalence of chatbots in human–computer communication has significantly increased. Emojis, as a form of emotional disclosure, have gained significant attention for their potential to boost chatbot service satisfaction. However, how and when emoji usage can increase satisfaction toward chatbots is not fully examined. This paper aims to fill this gap and contribute to the rapidly evolving field of human-chatbot communication research. Through three experiments, this paper investigates and explores the role of emojis in enhancing chatbot interactions. The results reveal that emojis heighten chatbot's perceived warmth but do not necessarily augment their competence. This warmth promoting effect leads to boosted service satisfaction and is more apparent when chatbots serve hedonic purposes and are pre-programmed rather than highly autonomous. However, the warmth upshot of emojis is not as potent for chatbots as it is for humans. While this study unravels the intricate pathway of how emojis augment service satisfaction, it also extends the dialogue of the Stereotype Content Model (SCM) and propels the new wave of the Computers Are Social Actors (CASA) paradigm. Thus, this research lays down pathways for further studies in understanding the role of emotionally simulated interactions in automated technologies.

The CASA theory no longer applies to desktop computers

The Computers Are Social Actors (CASA) theory is the most important theoretical contribution that has shaped the field of human–computer interaction. The theory states that humans interact with computers as if they are human, and is the cornerstone on which all social human–machine communication (e.g., chatbots, robots, virtual agents) are designed. However, the theory itself dates back to the early 1990s, and, since then, technology and its place in society has evolved and changed drastically. Here we show, via a direct replication of the original study, that participants no longer interact with desktop computers as if they are human. This suggests that the CASA Theory may only work for emergent technology, an important concept that needs to be taken into account when designing and researching human–computer interaction.

Gender Bias in Virtual Doctor Interactions: Gender Matching Effects of Chatbots and Users on Communication Satisfactions and Future Intentions to Use the Chatbot

A healthcare provider’s gender not only influences patients’ health-related behaviors but also the reliability of health information. Grounded in the Computers Are Social Actors (CASA) theory and prior health literature, the current study examines how healthcare chatbots’ gender cues and user gender affect intentions to use the chatbot and chatbot expertise perceptions. Using a 3 (Chatbot Gender Cues: Chatbot vs. Male Doctor vs. Female Doctor) X 2 (User Gender: Male vs. Female) between-subjects experiment, this study indicates that the female-doctor design cues led to significantly higher perceived warmth and communication satisfaction, which subsequently increased social presence and future intentions to use the chatbot. Results also indicated a significant gender congruence effect between female users and the female-doctor design cue chatbot to yield greater communication satisfaction. This study, however, did not find a significant difference in perceived expertise between male versus female doctor design cues. Theoretical and practical implications are discussed.

Minding the source: toward an integrative theory of human–machine communication

Abstract According to the computers are social actors (CASA) paradigm, a dominant theoretical framework for research on human–computer interaction, people treat computers as if they were people. Recent studies on human–machine communication (HMC) and human–artificial intelligence (AI) interaction, however, appear to focus on when and how people respond to machines differently than to human agents. To reconcile this apparent contradiction, this study reviews critically the two overarching theoretical explanations proposed and tested in each respective tradition, the mindlessness account and the machine heuristic. After elaborating on several conceptual and operational issues with each explanatory mechanism, an alternative theoretical model of HMC is proposed that integrates both research traditions and generates predictions that potentially deviate from the dual-process models. Lastly, it is discussed how recent developments in AI technology invite modifications to the current understanding of HMC and beyond.

SIoT robots and consumer experiences in retail: Unpacking repeat purchase intention drivers leveraging computers are social actors (CASA) paradigm

This study contributes to knowledge on the so far limited understanding of how to manage collaboration between Social Internet of Things (SIoT) service robots and consumers in the retail context. Embedded in Computers Are Social Actors (CASA) paradigm, we leverage a Partial Least Approach - Structural Equation Modelling (PLS-SEM) (n = 356) to show that word of mouth, consumer promotion experience, relationship quality, and inspiration significantly impact consumers’ repeat purchase intention when SIoT robots are used. Noticeably, while relationship quality is significant, it has a negative coefficient indicating that consumers may have high pre-existing anxiety towards SIoT service robots.

Artificial intelligence-generated virtual influencer: Examining the effects of emotional display on user engagement

Focusing on the application of artificial intelligence, this study investigates the impact of emotional display on user engagement with computer-generated imagery influencers through the lens of the computers are social actors (CASA) framework. It breaks down emotions into individual muscle movements (i.e., facial action units). By using facial recognition based on 1,028 pictures shared by Lil Miquela, the findings disclose the significance of happiness, sadness, disgust, and surprise in triggering user engagement when promoting diverse products with visually captivating content. The findings highlight the importance of balancing the intensity of muscle movement to streamline the interplay between technology, human behaviour, and digital communication.

“To comply or to react, that is the question:” the roles of humanness versus eeriness of AI-powered virtual influencers, loneliness, and threats to human identities in AI-driven digital transformation

AI-powered virtual influencers play a variety of roles in emerging media environments. To test the diffusion of AI-powered virtual influencers among social media users and to examine antecedents, mediators, and moderators relevant to compliance with and reactance to virtual influencers, data were collected using two cross-sectional surveys (∑ N = 1623). Drawing on the Diffusion of Innovations theory, survey data from Study 1 (N1 = 987) provide preliminary descriptive statistics about US social media users' levels of awareness of, knowledge of, exposure to, and engagement with virtual influencers. Drawing from the theoretical frameworks of the Uncanny Valley Hypothesis and the CASA (Computers Are Social Actors) paradigm, Study 2 examines social media users' compliance with versus reactance to AI-powered virtual influencers. Survey data from Study 2 (N2 = 636) provide inferential statistics supporting the moderated serial mediation model that proposes (1) empathy and engagement with AI-powered virtual influencers mediate the effects of perceived humanness versus eeriness of virtual influencers on social media users' behavioral intention to purchase the products recommended by the virtual influencers (serial and total mediation effects) and (2) loneliness moderates the effects of humanness versus eeriness on empathy. Drawing from the theory of Psychological Reactance, Study 2 further reports the moderation effect of social media users' trait reactance and perceived threats to one's own human identity on the relationship between perceived eeriness and compliance with versus situational reactance to virtual influencers. Theoretical contributions to CASA research and the Uncanny Valley literature as well as managerial implications for AI-driven digital transformation in media industries and virtual influencer marketing are discussed.

How to Provide Feedback? The Role of Robot’s Language and Feedback Framework

Approximately half of the world’s population are bilingual or multilingual. The foreign language effect has been a part of the academic study of people’s decision-making. The Computers are Social Actors (CASA) paradigm posited that people will respond differently when the robot used different social cues. The robot’s language might be an important social cue that makes people have different behavior and experience in the interaction. A limited amount of work has examined the effect of the robot’s language. Thus, this study investigated the effects of the robot’s language (native, foreign) and feedback framework (positive, neutral, negative) on users’ interaction with robots in cognitive tasks. This study further used fNIRS to investigate the effects of these factors on users’ brain activations. The results indicated that: (i) participants had marginally significantly higher performance, and felt significantly higher social presence when the robot used a foreign language; (ii) participants had significantly lower performance when the robot used a neutral feedback framework; (iii) participants had worse experience when the robot used a negative feedback framework; (iv) the use of a foreign language could reduce the effects of the robot’s feedback framework on participants’ performance and brain activations. These findings enrich the research on the foreign language effect, CASA paradigm and the feedback intervention. The findings also have implications for the design and development of robots and conversational agents.

Effect of AI Chatbot’s Interactivity on Consumers’ Negative Word-of-Mouth Intention: Mediating Role of Perceived Empathy and Anger

Although much research has examined chatbot interactivity as one of the most human-like qualities, replicating a conversation norm in human-to-human dialogue, far less has examined the "native state" of two-way communication between humans and chatbots. This study fills the gap by substantiating a synchronous two-way conversation between users and chatbots and testing how chatbot interactivity shapes the users’ perceived empathy. Guided by the computers-are-social-actors (CASA) paradigm, like skilled customer service representatives interacting with customers during product failures, this study designs a between-subject experiment (N = 300) to embody the warm and competent types of human conversation and suggests two versions of chatbot interactivities: social interactivity (high/low) and message interactivity (high/low). We further investigate whether a chatbot’s warm or competent persona manifested in its interaction with consumers elicits empathic responses from consumers, such as calming down their anger in line with social response theory. As a result, both social and message interactivity in chatbots effectively elicit users’ perception of empathy, alleviating customer anger and reducing adverse word-of-mouth (WOM) intentions. The findings provide empirical evidence that supports the explanatory framework of social response theory. This research demonstrates how technological features, particularly interactivity, contribute to creating experiential value for customers. Specifically, it elucidates how customers perceive empathy displayed by chatbots, which subsequently influences their level of anger and inclination to engage in negative word-of-mouth communication. Notably, these results offer novel insights into the roles of chatbot interactivity and empathy as psychological mechanisms that enable consumers to express their dissatisfaction by interacting with chatbot service agents, particularly in situations involving product failures. The implications of this study extend to the practical realm, providing valuable guidance on the design and implementation of empathetic chatbots in frontline consumer service contexts.

Investigating the impact of intelligent personal assistants on the purchase intentions of Generation Z consumers: The moderating role of brand credibility

Consumers increasingly use intelligent personal assistants for shopping, given their advantages of being hands-free and voice-controlled, enabling individuals to multi-task during interactions, and identifying users through voice printing. Members of Generation Z (Gen Z), the world's most populous generation, who focus on experiences and innovation and are familiar with using the internet, are increasingly using smart technologies. However, the factors that influence Gen Z consumers' purchase intentions as they use intelligent personal assistants (IPAs) are unknown. Drawing on the computers are social actors (CASA) paradigm, this paper explores how the intelligent features of IPAs affect Gen Z consumers' purchase intentions and investigates the moderating effect of brand credibility. Using structural equation modeling on a sample of 428 Gen Z consumers who use IPAs, the results show that personalization and a conversational tone have significant positive effects on informational support and emotional support. Autonomy and responsiveness have significant positive effects on informational support. Informational support and emotional support have a positive impact on purchase intentions. Brand credibility positively moderates the relationships between informational support and purchase intentions and between emotional support and purchase intentions. This study deepens our understanding of the antecedents of purchase intentions that affect Gen Z's use of IPAs and provides practical guidance for the development of IPAs. In addition, the results could help companies develop strategies to improve brand credibility to increase the purchase intentions of Gen Z consumers.

Sponsored by a robot? How the human likeness of virtual influencers influences purchase intentions towards sponsored products

With the rapid development of artificial intelligence (AI) technologies and the emerging trend of influencer marketing, virtual influencers have become promising advertising tools for marketers. Based on computers are social actors (CASA) and uncanny valley theory (UVT), this research study conducted a pilot study (N = 155) and an online experiment (N = 293) to investigate how the perceived humanness of virtual influencers affects their advertising effectiveness. The results indicate that the perceived humanness of virtual influencers increases perceptions of eeriness, attractiveness and authenticity. The increased perceived attractiveness and authenticity of virtual influencers, in turn, leads to more positive attitudes towards the virtual influencers and purchase intentions towards the advertised products. Furthermore, the current study found that such effects were moderated by the perceived similarity between consumers and virtual influencers. The theoretical and practical implications are discussed.

Does Self-Disclosing to a Robot Induce Liking for the Robot? Testing the Disclosure and Liking Hypotheses in Human–Robot Interaction

When someone intimately discloses themselves to a robot, does that make them like the robot more? Does a robot’s reciprocal disclosure contribute to a human’s liking of the robot? To explore whether these disclosure-liking effects in human–human interaction also apply to human–robot interaction, we conducted a between-subjects lab experiment to examine how self-disclosure intimacy (intimate vs. non-intimate) and reciprocal self-disclosure (yes vs. no) from the robot influence participants’ social perceptions (i.e., likability, trustworthiness, and social attraction) toward the robot. None of the disclosure-liking effects were confirmed by the results. In contrast, reciprocal self-disclosure from the robot increased liking in intimate self-disclosure but decreased liking in non-intimate self-disclosure, indicating a crossover interaction effect on likability. A post-hoc analysis was conducted to further understand these patterns. Implications in terms of the computers are social actors (CASA) paradigm were discussed.

Demystifying Computer-Generated Imagery (CGI) Influencers: The Effect of Perceived Anthropomorphism and Social Presence on Brand Outcomes

Computer-generated imagery (CGI) influencers are at the forefront of brand communication trends driven by artificial intelligence (AI). Focusing on Lil Miquela, this study explores the mechanism underlying the CGI influencer’s persuasiveness based on the computers-are-social-actors (CASA) framework. The study examines how consumers’ attitudes toward the influencer’s sponsored posts and the endorsed brands are impacted by her perceived anthropomorphism through mediators of social presence and perceived attractiveness. Based on a survey of generation Z consumers, we found that the CGI influencer’s perceived anthropomorphism effectively enhances her social presence, which in turn boosts her perceived physical and social attractiveness, respectively, to drive consumer evaluation outcomes.

Deep mind in social responses to technologies: A new approach to explaining the Computers are Social Actors phenomena

The Computers are Social Actors (CASA) paradigm has been increasingly used as a major theoretical framework to explain users' social responses to emerging technologies such as chatbots, voice assistants, and social robots. However, the core explanatory mechanism of the CASA paradigm is still under debate. In past works, mixed findings have emerged to support the mindlessness explanation and the mindfulness explanation. Thus, to better understand which mechanism features more explanatory power, this study analyzed 834 participants' responses and adopted a new approach combining the experimental design and classical multidimensional scaling. An investigation into participants' cognitive maps of technology differences in evoking social presence and perceived trustworthiness suggested that compared to mindfulness, mindlessness had more power in accounting for users’ social responses to technologies. The findings can serve as important evidence for the explanatory mechanism of the CASA paradigm, make methodological contributions, and have practical implications.