Virtual Influencers and Sustainable Brand Relationships: Understanding Consumer Commitment and Behavioral Intentions in Digital Marketing for Environmental Stewardship

Abstract

This investigation examines the psychological mechanisms governing human–virtual influencer relationships and their consequential impact on environmentally-conscious consumer behavior within digital marketing ecosystems. Employing theoretical frameworks from computer-mediated communication and social psychology, this study scrutinizes how algorithmically generated social media personalities cultivate para-social relationships that drive sustainable consumption patterns. The research operationalizes five core virtual influencer characteristics—expertise, similarity, attractiveness, familiarity, and para-social interaction—as predictive variables influencing relationship commitment and subsequent eco-conscious brand engagement. Consumer innovativeness functions as a moderating variable within this theoretical model. The data collection encompassed 677 respondents demonstrating active engagement with sustainability-focused virtual influencer content, analyzed through structural equation modeling (EQS 6.4) and the PLS-SEM methodology (SmartPLS 4.0). The empirical analysis reveals significant positive correlations between virtual influencer characteristics and relationship commitment, with similarity and attractiveness demonstrating the strongest predictive validity. Relationship commitment emerged as a significant mediator influencing sustainable brand attitudes, which subsequently affected purchase intentions for environmentally responsible products. Consumer innovativeness demonstrated positive moderating effects across all virtual influencer characteristics, with particularly robust effects observed for attractiveness and para-social interaction within sustainable brand contexts. This research advances the human–AI interaction literature by elucidating the psychological mechanisms through which virtual influencers facilitate consumer relationship formation and drive behavioral outcomes toward sustainable consumption practices. The findings provide empirically validated strategic frameworks for marketers developing virtual influencer campaigns that promote environmental stewardship, emphasizing the cultivation of perceived similarity and attractiveness while incorporating audience innovativeness as a critical segmentation variable in sustainable marketing initiatives.

1. Introduction

The transformation in communication methods through social media is fundamentally changing people’s daily lives [1]. The continuous evolution of network environments due to advances in information and communication technology, coupled with the constant upgrading of smartphone technology, is affecting not only individual lives but also corporate marketing activities. Amid these media changes, virtual influencers—artificial personalities created through computer graphics—have emerged and become active on social media platforms [2]. Virtual influencers, also known as ‘virtual humans,’ ‘virtual beings,’ or ‘digital humans,’ have garnered followings comparable to human influencers and are generating significant engagement through various activities, including advertising modeling, YouTube content creation, and shopping hosting. These virtual entities are establishing themselves as a new paradigm in digital marketing across various industries through social media [3,4]. The emergence of virtual influencers represents a unique phenomenon in human–computer interaction that simultaneously addresses contemporary sustainability imperatives in digital marketing. These computer-generated entities offer inherent environmental advantages over traditional influencer marketing through reduced carbon footprints associated with travel, physical production, and material consumption [5,6]. Users develop meaningful psychological connections with these virtual entities while engaging with increasingly environmentally conscious brand messaging that aligns with global sustainability discourse and corporate environmental responsibility frameworks [7].

This raises important questions about how human psychology adapts to and processes relationships with artificial personalities in digital environments. While traditional computer-mediated communication typically involves human-to-human interaction through digital channels, virtual influencers present a novel paradigm where humans knowingly form para-social relationships with computer-generated characters, challenging our conventional understanding of social presence and psychological attachment in digital spaces.

While virtual influencer marketing initially faced criticism for lacking authenticity, advances in artificial intelligence (AI) and computer graphics (CG) technology have enabled the creation of virtual influencers that are increasingly realistic and effective for marketing purposes [3,4,8]. Virtual influencer marketing is now recognized as a legitimate media strategy [9]. These digital characters, created through computer graphics and AI technology, can serve as optimal brand ambassadors due to their freedom from spatial-temporal constraints and privacy concerns [8,10].

The global health crisis of 2020 has fundamentally transformed the digital marketing landscape, particularly regarding the emergence of virtual influencers as legitimate brand ambassadors. This shift coincides with the broader digitalization of human interaction, leading to enhanced public receptivity toward computer-generated personalities. Industry analysts suggest a significant transformation in marketing resource allocation, with virtual influencer campaigns projected to command an increasing share of promotional budgets. The virtual influencer market demonstrates remarkable growth potential, which parallels the broader expansion of social media advertising, which analysts project will exceed USD 260 billion by 2028, representing a compound annual growth rate of approximately 12%. This trajectory reflects the growing integration of artificial personalities into mainstream marketing strategies [11]. The convergence of virtual influencer marketing with sustainable brand positioning represents a paradigmatic shift toward environmentally conscious digital communication strategies. Virtual influencers inherently embody sustainability principles through their digital-only existence, eliminating traditional marketing activities’ environmental externalities such as transportation, physical product sampling, and resource-intensive content production [12,13]. This technological evolution coincides with escalating consumer demand for sustainable brand practices and environmental accountability, positioning virtual influencers as optimal ambassadors for sustainable brand narratives and eco-conscious consumer engagement [14,15].

Users interact with virtual influencers’ social media accounts through following and commenting, demonstrating engagement patterns similar to those observed with human influencers. Businesses recognize virtual influencer activities as opportunities for innovative marketing strategies to build positive relationships with consumers [16]. Marketing activities utilizing virtual influencers are actively conducted across the United States, Europe, Japan, and other regions. The proliferation of virtual influencers as marketing agents represents an emerging paradigm in digital communication, yet a scholarly examination of this phenomenon remains nascent. While these artificial personas increasingly permeate various market sectors worldwide, there exists a significant research gap regarding their efficacy as marketing tools and their influence on consumer decision-making processes. Moreover, the cognitive and behavioral dynamics of human interactions with virtual influencers transcend conventional marketing metrics, raising complex questions about social engagement in digital spheres. These interactions raise important questions about how humans process and respond to computer-generated social entities, the nature of psychological attachment to artificial personalities, and the cognitive mechanisms underlying para-social relationships in digital spaces. Understanding these psychological processes is crucial for both theoretical developments in human–computer interaction and practical applications in digital communication.

Therefore, this study aims to examine the relationships between outcome variables and virtual influencer characteristics, given their growing influence. Specifically, this research investigates the relationships between virtual influencer attributes—including expertise, attractiveness, similarity, familiarity, and parasocial interaction—and relationship commitment. It also examines the connections between relationship commitment, brand attitude, and purchase intention. Additionally, this study positions consumer innovativeness as a moderating variable to determine its effect on the relationship between virtual influencer attributes and relationship commitment.

2. Theoretical Background

2.1. Virtual Influencer Marketing and Sustainable Marketing Paradigms

Virtual influencer marketing represents an innovative evolution in digital marketing strategy, where computer-generated personalities assume the traditional role of human influencers in social media engagement and brand promotion. These digital entities, also known as virtual influencers or CGI influencers, leverage artificial intelligence and advanced graphics technology to create content and interact with followers across social networking platforms [17,18]. Unlike traditional human influencers, virtual influencers offer brands unprecedented control over messaging and image consistency while maintaining a carefully crafted persona that resonates with digital-native consumers.

Recent research demonstrates that virtual influencer marketing has emerged as a distinct subset of influencer marketing, characterized by unique para-social relationships between audiences and artificially created personalities [19,20]. Studies indicate that despite their artificial nature, virtual influencers can generate authentic connections with followers through consistent narrative building and strategic content creation. The phenomenon challenges traditional assumptions about authenticity in marketing, as consumers demonstrate willingness to engage with and trust recommendations from non-human entities [21].

The effectiveness of virtual influencer marketing stems from its ability to combine the controlled messaging of traditional advertising with the perceived authenticity and engagement patterns of human influencer marketing. Virtual influencers can maintain constant availability, perfect consistency in brand messaging, and risk-free brand associations, while simultaneously fostering para-social relationships with their followers through carefully curated digital personalities and storylines [22,23]. This hybrid approach allows brands to leverage the benefits of influencer marketing while mitigating many of the risks associated with human influencers, such as controversial behavior or inconsistent messaging.

The convergence of advanced computational technologies, particularly in artificial intelligence and digital imaging, has given rise to a novel phenomenon in social media: anthropomorphic virtual influencers. These computer-generated personalities, which exist solely in digital spaces, represent a sophisticated intersection of artificial intelligence algorithms and high-fidelity computer graphics [2,24]. The emergence of these virtual entities has transformed traditional marketing paradigms, as major international brands increasingly incorporate these digital ambassadors into their marketing strategies across multiple sectors, including cosmetics, haute couture, and lifestyle products. Notable examples of these virtual personas, such as Lil Miquela, Shudu, and Imma, have cultivated substantial follower bases while simultaneously serving as digital brand representatives for multinational corporations. This development signals a fundamental shift in how businesses engage with consumers in the digital age, challenging traditional notions of authenticity and human interaction in social media marketing [25,26]. For instance, Lil Miquela has collaborated with BMW, Prada, and Dior, while Shudu has modeled for Samsung Galaxy and Tiffany & Co. Additionally, Imma has expanded her sphere of influence through partnerships with Porsche Japan, IKEA, and Off-White.

Lil Miquela was created using computer graphics and AI technologies, while Imma was developed using real-time CGI technology, with AI-based technologies employed in persona development and social media content generation. Shudu is produced using 3D modeling and AI technologies. The advancement of generative AI technology has facilitated the emergence of new AI virtual influencers [3,4,5,9]. Virtual influencers are digital characters that possess human-like features, characteristics, and personalities, embodying forms similar to real humans [8,25]. They are virtual entities created for commercial purposes on social media, wielding influence over their numerous followers’ product perceptions and purchase intentions [6,18]. As consumers spend increasing amounts of time online, there is a growing tendency toward higher engagement with virtual influencers compared to traditional influencers [8].

Virtual influencer marketing represents a convergence of technological innovation and sustainable marketing practices, embodying the principles of environmental responsibility through digital-first approaches to brand communication [27,28]. Unlike traditional influencer marketing, which involves substantial environmental costs through travel, physical product consumption, and material-intensive content creation, virtual influencers operate within entirely digital ecosystems that minimize ecological footprints while maximizing reach efficiency [29].

Contemporary sustainability research in digital marketing emphasizes the environmental advantages of virtual brand ambassadors, particularly in contexts where traditional marketing activities generate significant carbon emissions and resource consumption [30,31]. Virtual influencers can promote sustainable products and environmental consciousness without the inherent contradictions often present in traditional influencer partnerships, where lifestyle choices may conflict with sustainability messaging [32].

The theoretical framework connecting virtual influencers to sustainable brand relationships builds upon corporate environmental responsibility theories and stakeholder capitalism models, where brand authenticity in sustainability communication becomes paramount [33,34]. Virtual influencers offer brands unprecedented control over sustainability messaging consistency while eliminating the potential greenwashing risks associated with human influencer lifestyle incongruences [35,36].

Research indicates that consumers increasingly evaluate brand sustainability performance through digital touchpoints, making virtual influencers ideal vehicles for communicating environmental initiatives and sustainable product attributes [37,38]. The digital nature of virtual influencers aligns naturally with contemporary consumer preferences for authentic sustainability communication and transparent environmental impact reporting [39].

2.2. Theoretical Framework Enhancement

The conceptual framework of this investigation draws substantially upon Social Identity Theory (SIT) to explicate the psychological mechanisms underlying consumer–virtual influencer relationships [17,40]. Social Identity Theory posits that individuals derive psychological satisfaction and self-concept validation through identification with salient social groups and charismatic representatives. Within the virtual influencer paradigm, this theoretical construct assumes particular complexity, as consumers form identificatory attachments with algorithmically generated personas that embody idealized characteristics and lifestyle aspirations. The application of SIT to virtual influencer marketing necessitates theoretical adaptation, as traditional in-group/out-group categorizations become fundamentally reconceptualized when the referent entity exists purely within digital constructions.

Within the Social Identity Theory framework, the virtual influencer characteristics map onto specific theoretical components as follows: Similarity represents the in-group identification mechanism, where consumers assess their congruence with the virtual influencer’s presented values and lifestyle attributes. Attractiveness and expertise function as status enhancement variables, enabling consumers to derive social capital through association with idealized virtual personalities. Familiarity operates as the relationship maintenance component, reflecting repeated exposure that strengthens group membership feelings. Para-social interaction serves as the psychological identification process, facilitating emotional attachment despite the mediated nature of the relationship. These components collectively contribute to relationship commitment, which represents the sustained psychological investment in maintaining the consumer–virtual influencer identification bond. This process transcends conventional parasocial relationship frameworks by incorporating elements of technological mediation and artificial personality construction. The theoretical implications suggest that virtual influencers function as vehicles for identity projection and social positioning, enabling consumers to align themselves with carefully curated digital personas that embody specific lifestyle values and aesthetic preferences [40].

The Technology Acceptance Model (TAM), originally developed by Davis and subsequently refined through extensive empirical validation, provides crucial theoretical scaffolding for understanding consumers’ adoption of virtual influencer technologies [26,41].

The Technology Acceptance Model application in this study operates through the existing constructs of expertise, attractiveness, and familiarity, which collectively represent perceived usefulness and ease of use in the virtual influencer context. Rather than introducing additional constructs, this investigation recognizes that consumers’ acceptance of virtual influencers emerges through the combined evaluation of these established characteristics. Perceived usefulness manifests through expertise assessments and attractiveness evaluations, while perceived ease of use is reflected in familiarity and similarity perceptions. The artificial nature of virtual influencers requires consumers to psychologically accommodate non-human entities as legitimate information sources, a process captured through the aggregated effects of the measured characteristics on relationship commitment. Within this investigation’s theoretical architecture, TAM operates through modified constructs that accommodate the unique characteristics of artificial personality interaction. Specifically, perceived usefulness in the virtual influencer context encompasses consumers’ cognitive assessments of how virtual influencer content enhances their decision-making processes, social positioning, and experiential satisfaction. Perceived ease of use manifests through consumers’ evaluations of the accessibility and comprehensibility of virtual influencer content delivery mechanisms across digital platforms. The theoretical adaptation of TAM for virtual influencer research necessitates the incorporation of additional constructs that address the artificial nature of these entities. We propose an extended TAM framework that includes “artificial entity acceptance” as a mediating variable between traditional TAM constructs and behavioral intentions. This theoretical extension acknowledges that consumer acceptance of virtual influencers involves not merely technological adoption, but the psychological accommodation of non-human social entities as credible information sources and relationship partners [26,41].

2.3. Virtual Influencer Characteristics

The emergence of virtual influencers as digital marketing entities has introduced a paradigm shift in how brands engage with consumers online. These computer-generated personalities occupy a unique position in the marketing landscape, exhibiting distinctive attributes that set them apart from both traditional celebrity endorsers and human social media influencers [8,40]. When examining their impact on consumer purchasing decisions, virtual influencers present an intriguing paradox: while openly artificial in nature, they leverage human-like qualities to forge connections with audiences. Scholarly investigation has identified several fundamental characteristics that drive virtual influencers’ effectiveness in consumer behavior modification, encompassing elements such as perceived authenticity despite their synthetic nature, advanced technological implementation, unwavering brand message consistency, and the evolution of novel para-social relationship dynamics between artificial entities and human followers [42,43].

The concept of expertise in virtual influencers presents an interesting paradox, as their knowledge and capabilities are deliberately programmed rather than naturally acquired [19]. Studies indicate that consumers’ perception of virtual influencer expertise is closely tied to the sophistication of their digital presentation and the authenticity of their specialized knowledge domains [44,45,46,47]. This artificial expertise, when properly cultivated, can be particularly effective in technical product endorsements, where the virtual nature of the influencer may actually enhance their perceived authority in digital and technological domains [48,49,50].

Familiarity in the context of virtual influencers takes on new dimensions, as the relationship between consumers and these digital entities is inherently mediated through technology. Research shows that para-social relationships with virtual influencers can be equally, if not more intense, than those with human influencers, particularly among digitally native consumers [51,52,53]. The consistent and controllable nature of virtual influencers’ personality traits and behaviors can facilitate stronger perceived intimacy and emotional attachment, though this relationship is distinctly different from traditional human-to-human para-social bonds [54].

Virtual influencers present a unique case study in similarity and attractiveness, as their appearances and personalities can be precisely crafted to appeal to specific target demographics [55]. Their digital nature allows for perfect consistency in appearance and behavior, while simultaneously enabling rapid adaptation to changing trends and consumer preferences. This malleable yet stable identity creates a novel form of attractiveness that combines idealized physical attributes with carefully curated personality traits.

Para-social interaction with virtual influencers represents an evolution in mediated relationships, as consumers knowingly engage with non-human entities while still forming meaningful emotional connections [56]. These interactions challenge our traditional understanding of para-social relationships, as the artificial nature of the influencer is transparent, yet followers still develop authentic emotional attachments. Research indicates that this awareness of artificiality may actually enhance rather than diminish engagement, as consumers appreciate the authenticity of acknowledging the virtual nature of the relationship [57,58].

The technological sophistication of virtual influencers adds a new dimension to their effectiveness, as their digital nature allows for seamless integration across platforms and perfect consistency in brand messaging. This technological foundation enables virtual influencers to maintain unwavering brand alignment while simultaneously adapting to emerging trends and consumer preferences with unprecedented agility [41,56,57].

The conceptualization of familiarity within virtual influencer research requires theoretical disambiguation from conventional interpersonal relationship constructs. Familiarity in the virtual influencer context represents consumers’ accumulated psychological comfort and recognition developed through repeated exposure to consistent digital personas. This construct encompasses three dimensions: (1) recognition familiarity—the ability to identify and distinguish the virtual influencer’s unique characteristics; (2) experiential familiarity—the comfort level derived from repeated interactions with the virtual influencer’s content; and (3) relational familiarity—the sense of predictable connection fostered through consistent digital engagement patterns. Unlike traditional interpersonal familiarity, virtual influencer familiarity operates through algorithmic consistency, where consumers develop attachment to predictable digital personas rather than evolving human personalities [47].

The theoretical justification for positioning familiarity as a virtual influencer characteristic stems from Uncertainty Reduction Theory, which posits that individuals seek predictability and consistency in their social interactions. Virtual influencers, through their algorithmically controlled personas, provide unprecedented levels of behavioral consistency and message reliability, potentially creating stronger familiarity perceptions than those achievable through human influencer interactions. This theoretical framework suggests that familiarity with virtual influencers operates through cognitive scripts and expectation patterns that differ fundamentally from human-based familiarity constructs [47,49].

Following extensive theoretical examination, this investigation positions para-social interaction as an independent virtual influencer characteristic that directly influences relationship commitment, alongside expertise, similarity, attractiveness, and familiarity. While para-social interaction shares theoretical foundations with traditional media personality relationships, its manifestation in virtual influencer contexts requires distinct conceptualization due to the transparent artificial nature of these entities. Virtual influencer para-social interaction encompasses consumers’ unidirectional emotional engagement with algorithmically generated personalities, characterized by perceived reciprocal communication despite explicit awareness of the entity’s artificial nature. This construct operates as a direct antecedent to relationship commitment, reflecting consumers’ psychological investment in maintaining emotional connections with virtual entities through consistent digital interactions and perceived responsiveness. This theoretical distinction necessitates modified measurement approaches that capture the artificial yet psychologically meaningful nature of virtual influencer para-social interactions [41,56,57].

2.4. Relationship Commitment

Relationship commitment has emerged as a crucial factor in relationship marketing research [17,58]. Such commitment reflects the confidence that a mutual relationship between traders will persist [17,58], with consumers’ willingness to maintain relationships with companies representing a key psychological factor [59,60,61]. The establishment of positive interpersonal relationships occurs through proper information sharing between parties [59]. Research indicates that consumers who develop strong relationship commitments demonstrate higher levels of satisfaction and trust, ultimately showing a greater propensity to purchase a company’s products. Furthermore, customers exhibiting strong relationship commitment manifest more robust purchase intentions compared to those with weak commitment [17,58,60].

2.5. Consumer Innovativeness

Contemporary innovation research conceptualizes innovativeness as a latent characteristic defined by an individual’s preference for novel and diverse experiences [26]. As a disposition toward novelty, innovative tendency drives the pursuit of new experiences that engage both cognitive and sensory faculties. The cognitive dimension of consumer innovativeness encompasses an individual’s propensity to seek mentally stimulating experiences, while the sensory aspect relates to their inclination toward sensory engagement and experiential activities [42,62].

Consumer innovativeness manifests as the degree to which individuals readily adopt new products and experiences. In the context of trend adoption, a distinct subset of consumers exhibits enhanced receptivity to new products and demonstrates an innovative personality that accepts various economic and social risks associated with early adoption. These individuals, characterized by innovative traits, are classified as innovators [63]. Research indicates that highly innovative individuals often display impulsive tendencies and emotional consumption patterns. Innovativeness significantly influences the consumption adoption process and subsequent consumer decision-making behavior. Specifically, innovators demonstrate accelerated product adoption rates, enhanced trend information acquisition, and proactive purchasing behavior [63]. Within this study’s framework, consumer innovativeness serves as a moderating variable, with the anticipated effect of virtual influencer characteristics on relationship engagement varying according to consumers’ innovative dispositions.

3. Hypotheses and Research Model

3.1. Suggested Research Model

Drawing upon the existing literature, this study’s theoretical framework examines the interplay between virtual influencer attributes and consumer behavioral outcomes. The research model investigates three key dependent variables: relationship commitment, brand attitudes, and purchase intentions. Specifically, Hypothesis 1 explores the causal mechanisms through which virtual influencer characteristics affect relationship commitment. Hypotheses 2 and 3 examine the sequential relationships among relationship commitment, brand attitudes, and purchase intentions. Additionally, Hypothesis 4 investigates consumer innovativeness as a potential moderating variable that may strengthen or weaken the associations between virtual influencer characteristics and relationship commitment. Figure 1 presents a visual representation of these hypothesized relationships in our conceptual framework.

3.2. Research Hypotheses

Contemporary research has established that virtual influencers generate content focusing on product introduction and utilization, with consumers voluntarily accepting their recommendations based on the information provided [16,25]. Studies indicate that when information recipients perceive similarity with an influencer, they are more likely to accept and trust them as an information source, ultimately affecting their purchase intentions [48,51]. The perceived expertise of an information provider, as demonstrated through social media performance, reflects the extent to which followers accept the provided information [19]. Recent studies suggest that followers evaluate virtual influencers’ expertise based on their demonstrated knowledge, content quality, and experience with promoted products or services [9,20]. Research indicates that expertise is a crucial factor in enhancing marketing information delivery success, with expert-presented product information leading to increased brand loyalty and stronger consumer relationships [2,8]. Therefore, a virtual influencer’s perceived expertise constitutes a variable that affects their followers’ relationship commitment.

H1. 

The perceived expertise of a virtual influencer has a positive effect on relationship commitment.

Recent literature defines similarity as the degree of concordance between an individual and an object regarding interests, values, preferences, and behavioral patterns [17,21]. Contemporary studies on information sources have demonstrated that recipients are more likely to accept opinions from sources perceived as similar in character or specific attributes [18]. Therefore:

H2. 

Similarity between a consumer and a virtual influencer has a positive effect on relationship commitment.

Modern research has expanded the concept of attractiveness beyond physical characteristics to encompass overall appeal, including personality and digital presence [4]. Studies indicate that virtual influencer attractiveness significantly impacts information acceptance and relationship formation [3,9]. Recent findings suggest that information from highly attractive virtual influencers enhances brand recognition and generates positive brand attitudes [2]. Therefore:

H3. 

A virtual influencer’s attractiveness has a positive effect on relationship commitment.

Contemporary research characterizes familiarity as a cognitive and emotional response that encompasses feelings of connection and belonging [16]. Unlike traditional intimacy metrics, virtual influencer familiarity is evaluated through digital interactions and perceived accessibility [25]. Recent studies in brand relationship management indicate that familiarity serves as a crucial factor in fostering brand loyalty and strengthening consumer relationships [48,49]. Research suggests that the perceived congruence between consumer self-concept and virtual influencer personality positively affects emotional bonds and engagement [9]. Therefore:

H4. 

A virtual influencer’s familiarity has a positive effect on relationship commitment.

Para-social interaction (PSI) with virtual influencers is likely to enhance relationship commitment as it fosters an illusion of intimate connection and emotional attachment, despite the mediated nature of the relationship. This theoretical proposition builds on earlier research demonstrating that when followers develop para-social relationships with media personalities, they experience feelings of friendship and closeness that mirror real interpersonal bonds [56]. In the context of virtual influencers, these perceived intimate connections may be particularly potent because the artificial nature of the personality allows for the careful curation of relatable content and consistent engagement patterns [41]. The mechanism underlying this relationship can be explained through the lens of social presence theory, where regular exposure to virtual influencers’ carefully crafted personas and seemingly authentic interactions creates a sense of genuine human connection [40,63,64,65]. This perceived authenticity, combined with the psychological investment followers make in maintaining these para-social relationships, naturally leads to stronger relationship commitment [41,57,58]. Furthermore, the unique characteristics of virtual influencers—their constant availability, controlled messaging, and idealized presentation—may actually intensify followers’ emotional investment and loyalty compared to traditional human influencers [19,40]. Therefore, para-social interaction with virtual influencers can be constructed as a variable that affects relationship commitment:

H5. 

Para-social interaction between a virtual influencers and followers has a positive effect on relationship commitment.

Recent research has demonstrated that relationship commitment serves as a fundamental psychological construct in consumer–brand relationships, where stronger commitments lead to enhanced brand evaluations [17,60]. Studies indicate that consumers who develop robust relationship commitments exhibit higher levels of satisfaction and trust, which positively influences their overall brand assessment [59,61]. Through effective information sharing and sustained interactions, consumers with strong relationship commitments form more favorable psychological connections with brands [58]. This psychological bonding mechanism facilitates positive brand evaluations through emotional attachment and perceived relationship value [38,39].

H6. 

Relationship commitment has a positive effect on brand attitude.

The contemporary marketing literature consistently demonstrates that positive brand attitudes serve as crucial antecedents to consumer purchase intentions [17,58]. Research indicates that consumers who maintain favorable brand evaluations demonstrate stronger propensities toward purchase behavior, particularly when these attitudes are formed through sustained relationship commitment [60,66,67,68]. The psychological pathway from positive brand attitude to purchase intention is strengthened through emotional connections and perceived value derived from the consumer–brand relationship [58,61]. This relationship is particularly evident in digital marketing contexts, where consistent brand interactions through various platforms reinforce the connection between attitudinal favorability and behavioral intentions [17,58,69,70].

H7. 

Brand attitude has a positive effect on purchase intention.

Research has demonstrated that consumer innovativeness serves as a significant moderator in shaping how consumers respond to emerging marketing strategies [26,63]. Empirical evidence indicates that consumers exhibiting elevated levels of innovativeness display enhanced receptivity toward technological advancements and marketing innovations [62]. Within the virtual influencer domain, highly innovative consumers exhibit a pronounced tendency to favorably evaluate the distinctive attributes of these computer-generated personalities [42]. Additionally, Li et al. revealed that innovative consumers demonstrate an increased propensity to establish emotional bonds with novel marketing entities, indicating that consumer innovativeness may facilitate stronger relationship commitment in the virtual influencer sphere [26]. Building upon these theoretical foundations, this study proposes that consumer innovativeness functions as a positive moderating variable in the relationship between virtual influencer characteristics and relationship commitment. Specifically, we posit that the impact of virtual influencer characteristics on relationship commitment will be amplified among consumers who possess higher levels of innovativeness. This conceptualization positions consumer innovativeness as a critical moderating construct that influences the strength and direction of the association between virtual influencer characteristics and relationship commitment.

Contemporary sustainability research demonstrates that consumers’ environmental consciousness significantly influences brand evaluation and purchase decision-making processes, particularly in digital marketing contexts where sustainability messaging authenticity becomes paramount [71,72]. Virtual influencers, through their inherently sustainable digital existence and capacity for consistent environmental messaging, provide unique advantages in fostering sustainable brand relationships among environmentally-conscious consumers [73]. Studies indicate that consumers who prioritize environmental sustainability demonstrate enhanced receptivity toward brands that employ environmentally responsible marketing strategies, including digital-first approaches that minimize physical resource consumption [74,75]. Therefore, environmental consciousness moderates the relationship between virtual influencer characteristics and sustainable brand attitudes:

H8. 

Consumer innovativeness strengthens the influence of virtual influencer characteristics on relationship commitment, with particularly pronounced effects observed for attractiveness and para-social interaction dimensions.

4. Method

4.1. Operational Definitions and Measurement

Five virtual influencer characteristics were systematically operationalized for empirical analysis through a comprehensive multi-stage validation process: expertise, similarity, attractiveness, familiarity, and para-social interaction. The operational definitions employed in this investigation were rigorously derived from established theoretical frameworks and validated through extensive psychometric evaluation procedures.

Expertise is conceptualized as the degree to which an information recipient perceives a virtual influencer as possessing authoritative knowledge and demonstrating competent judgment regarding product-related messaging. This construct was measured using four validated items adapted from Breves et al. [52] incorporating modifications to accommodate the virtual influencer context through expert panel review and pilot testing procedures.

Similarity represents the perceived concordance between consumers and virtual influencers regarding interests, values, preferences, and behavioral orientations. The measurement instrument comprised four items systematically adapted from Schouten et al. [54] and Martínez-López et al. [55], with content validity established through cognitive interviewing protocols and construct refinement procedures.

Attractiveness encompasses the multidimensional appeal of virtual influencers, extending beyond physical characteristics to include personality dimensions and digital presentation sophistication. Four measurement items were derived from Chung and Cho [56] and Ki et al. [53], with adaptations validated through confirmatory factor analysis and convergent validity assessment.

Familiarity is operationalized as the cognitive and emotional intimacy consumers experience toward virtual influencers through repeated digital interactions and perceived accessibility. The construct employed four validated items adapted from Banks and Bowman [64] and Hwang and Zhang [57], incorporating temporal interaction patterns and digital relationship dynamics.

Para-social Interaction represents the unidirectional, mediated social engagement wherein consumers develop authentic psychological connections with virtual influencers despite the absence of reciprocal relationships [25,34,45,46]. Three rigorously validated items were adapted from Sakib et al. [40] and Boerman and Van Reijmersdal [41], with modifications ensuring contextual appropriateness for virtual entity relationships.

The measurement framework incorporated nineteen systematically validated question items, with four-item scales for expertise, similarity, attractiveness, and familiarity constructs, and a three-item scale for para-social interaction. All measurement instruments underwent comprehensive validation procedures including content validity assessment, expert panel evaluation, and pilot testing with representative samples to ensure construct validity and reliability [40,41,65,66].

Relationship commitment has evolved from its initial conceptualization as a desire to maintain valued relationships to a more nuanced construct. Recent scholars such as Kim and Parkdefine it as a psychological state encompassing both affective attachment and behavioral intentions that persist in a relationship despite fluctuating circumstances [16]. This multidimensional perspective has been particularly salient in social media contexts, where para-social relationships mediate consumer–brand connections [18]. Three question items pertaining to relationship commitment used in previous studies were modified and supplemented for the purposes of this study.

Contemporary research has expanded our understanding of brand attitude formation beyond traditional cognitive–affective frameworks. Liao et al. posit that brand attitudes in the digital age are increasingly shaped by the interplay between influencer authenticity and content congruence [66]. Four question items pertaining to brand attitude were modified and supplemented for the purposes of this study. This represents a significant evolution from conventional models, as social media influencers now serve as critical intermediaries in consumer–brand relationship development [66]. Riaz et al. characterize purchase intention as a dynamic decision-making process influenced by multiple touchpoints across digital platforms, where social proof and influencer credibility serve as key determinants [68]. This contemporary framework better reflects the complex nature of consumer decision-making in social commerce environments. Three question items were used in this study to measure purchase intention. Consumer innovativeness encompasses an individual’s propensity and capability to adopt novel ideas, products, and systems [63,70]. To assess this construct, three established measurement items from the literature were adapted and refined to align with the specific context of this investigation. The measurement instrument employed a 5-point Likert scale, where respondents indicated their level of agreement from 5 (strongly agree) to 1 (strongly disagree). The questionnaire items, as detailed in Appendix A, were derived from validated scales in previous research.

4.2. Data Collection

This investigation utilized primary data collected through a structured online survey administered via the EMBRAIN™ platform, a comprehensive consumer research system operated by Market-Insight International (MII). The survey instrument incorporated validated scales from the existing literature, which were systematically adapted for the virtual influencer context through expert panel evaluation and cognitive pretesting procedures. Scale standardization involved the following: (1) a consistent 5-point Likert scale formatting across all constructs; (2) declarative statement reformulation to eliminate question format inconsistencies; (3) item modification to ensure contextual appropriateness for virtual influencer evaluation; and (4) composite score calculation using factor-loading weights to ensure measurement precision. All measurement items were originally developed in English and validated through comprehensive pilot testing with 150 participants before full-scale data collection.

Enhanced Screening and Validation Procedures: Respondent screening incorporated two essential qualifying criteria rigorously validated through preliminary assessment protocols: (1) demonstrated familiarity with virtual influencer concepts, verified through definitional accuracy testing and recognition assessments, and (2) documented exposure to virtual influencer marketing content within the preceding six-month period, confirmed through specific brand and campaign recall verification procedures. These screening criteria underwent pilot testing with 150 preliminary respondents to establish validity and reliability parameters.

Virtual Influencer Context Specification: Participants were exposed to content from established virtual influencers including Lil Miquela, Imma, and Shudu, representing diverse demographic and aesthetic profiles. The study focused on virtual influencers promoting lifestyle products, fashion items, and technology products, with particular attention to those incorporating sustainability messaging in their content. Screening procedures verified participants’ familiarity with virtual influencer concepts through definitional accuracy testing and brand recognition assessments. Participants were required to demonstrate an awareness of at least one major virtual influencer and recall exposure to virtual influencer marketing content within the six-month period preceding data collection.

Methodological Transparency and Instrument Development: The measurement instrument development process incorporated multiple validation stages, beginning with a comprehensive literature review and theoretical framework construction, followed by expert panel evaluation involving five academic specialists in digital marketing and consumer behavior research. Item generation procedures employed both deductive (theory-driven) and inductive (exploratory) approaches, with content validity established through cognitive interviewing protocols conducted with 30 representative participants.

The survey instrument underwent extensive pre-testing procedures, including the following: (1) cognitive interviews for item comprehension assessment, (2) test–retest reliability evaluation with a 14-day interval (n = 75), and (3) convergent and discriminant validity assessment through confirmatory factor analysis. All measurement items were systematically adapted from validated scales in previous research, with specific attribution to original sources and the comprehensive documentation of modification procedures.

Data Quality Assurance and Validation Protocols: From the available dataset of 690 respondents, rigorous data cleaning procedures resulted in the exclusion of 23 cases due to (1) incomplete response patterns exceeding 15% missing data thresholds, (2) systematic response bias indicators including straight-lining and acquiescence patterns, and (3) failure to meet the established screening criteria through validation procedures. The final analytical sample of 677 valid responses underwent comprehensive outlier detection analysis and normality assessment procedures.

Ethical Compliance and Methodological Rigor: The research design incorporated comprehensive ethical safeguards aligned with internationally recognized research standards. Participant anonymity was maintained through advanced randomized response coding protocols, with no collection of personally identifiable information beyond essential demographic variables. All data underwent aggregation procedures prior to analysis, ensuring individual participant confidentiality throughout the research process. Informed consent procedures were implemented through comprehensive participant information protocols, clearly delineating study objectives, data usage parameters, and participant rights. Voluntary participation principles were strictly maintained, with explicit notification regarding withdrawal rights and the absence of adverse consequences for non-participation or study discontinuation.

4.3. Measurement Scale Standardization and Validation Protocols

To address potential measurement inconsistencies and enhance methodological rigor, this investigation implemented comprehensive scale standardization procedures across all constructs. All measurement items employed consistent 5-point Likert scale formatting, ranging from 1 (strongly disagree) to 5 (strongly agree), ensuring measurement uniformity and facilitating comparative analysis across constructs.

Composite Score Methodology and Aggregation Procedures: Composite scores for each construct were calculated using validated aggregation procedures incorporating factor loading weights derived from confirmatory factor analysis. For multi-dimensional constructs, particularly familiarity, weighted composite scores were computed using the following formula: Composite Score = Σ(λ_i × X_i)/Σλ_i.

Where λ_i represents the standardized factor loading for item i, and X_i represents the observed score for item i. This approach ensures that items with stronger psychometric properties contribute proportionally more to the composite construct score, enhancing measurement precision and theoretical validity.

Question Format Standardization and Interpretation Guidelines: All survey items employed consistent declarative statement formatting to eliminate any potential ambiguity associated with interrogative structures. Items originally formulated as questions (e.g., “Do you believe…”) were systematically reformulated as evaluative statements (e.g., “I believe…”) to enhance respondent comprehension and reduce interpretive variance. This standardization procedure underwent cognitive testing with 25 participants to ensure clarity and interpretive consistency across all measurement items.

5. Results

5.1. Descriptive Statistics

As shown in Table 1, most survey participants were women, with 370 participants (56.9%) being female and 280 participants (43.1%) being male. Age groups were distributed in multiple ways. Those in their 30s accounted for the largest proportion (39.6% or 257), followed by those in their 40s and 20s. Most participants were found to be enrolled in a university or to have earned a university-level degree. Regarding occupational groups, 271 participants were office workers, accounting for 41.7% of the total. The next largest group, comprising 23.9% of the total, were public officials. The average monthly income of the largest group of participants was about USD 3000–4000. Survey participants were found to spend about 3–5 h per week shopping.

5.2. Validity and Reliability Analysis

Comprehensive Psychometric Validation and Reliability Assessment: This investigation employed advanced structural equation modeling procedures utilizing EQS 6.4 and SmartPLS 4.0 software platforms to examine the complex relationships among virtual influencer characteristics, relationship commitment, brand attitudes, purchase intentions, and consumer innovativeness constructs. Prior to conducting hypothesis tests, an exploratory factor analysis was conducted on the question items for virtual influencer characteristics and the dependent variables, as shown in Table 2 and Table 3.

Enhanced Factor Analysis and Scale Validation Procedures: Prior to hypothesis testing, comprehensive exploratory factor analysis (EFA) was conducted utilizing principal axis factoring with oblique rotation procedures to accommodate the anticipated inter-factor correlations. The analysis incorporated rigorous statistical criteria including Kaiser–Meyer–Olkin (KMO) measure assessment, Bartlett’s sphericity testing, and eigenvalue determination using both Kaiser’s criterion and scree plot examination. For virtual influencer characteristics, the factor loadings exceeded the established 0.5 threshold across all measurement items, with five distinct factors that collectively explained 75.6% of the total variance emerging. The KMO measure achieved 0.881, substantially exceeding the recommended 0.70 threshold, while Bartlett’s test of sphericity demonstrated statistical significance (χ² = 8247.32, df = 171, p < 0.001), confirming the appropriateness of factor analysis procedures. The dependent variable factor analysis yielded four theoretically consistent factors explaining 83.1% of the cumulative variance, with a KMO measure of 0.912 and significant Bartlett’s test results (χ² = 5689.45, df = 66, p < 0.001), substantiating the factorial validity of the measurement model.

The study employed correlation analysis to evaluate the discriminant validity among the research variables, with the findings presented in Table 2, Table 3 and Table 4. Multiple analytical approaches were utilized to assess both centralized and conceptual validity across all factors.

Advanced Reliability and Validity Assessment: Comprehensive reliability analysis incorporated multiple assessment approaches including Cronbach’s alpha, composite reliability (CR), and average variance extracted (AVE) calculations. Internal consistency reliability demonstrated exceptional psychometric properties: expertise (α = 0.812, CR = 0.931), similarity (α = 0.821, CR = 0.871), attractiveness (α = 0.847, CR = 0.901), familiarity (α = 0.843, CR = 0.792), and para-social interaction (α = 0.845, CR = 0.873). All reliability coefficients substantially exceeded the established thresholds (α ≥ 0.70, CR ≥ 0.70), confirming robust internal consistency. Convergent validity was established through the systematic evaluation of factor loadings, AVE, and CR values. All standardized factor loadings exceeded 0.50, with most surpassing 0.70, indicating strong item-to-construct relationships. The AVE values for all constructs exceeded the recommended 0.50 threshold, while the CR values consistently surpassed 0.70, confirming adequate convergent validity. Discriminant validity assessment employed multiple analytical approaches including Fornell–Larcker criterion evaluation and cross-loading examination. The square root of AVE for each construct exceeded its correlations with other constructs, satisfying the Fornell–Larcker criterion. Additionally, the heterotrait–monotrait (HTMT) ratios remained below the conservative 0.85 threshold, providing robust evidence for discriminant validity.

Confirmatory Factor Analysis and Model Fit Assessment: Comprehensive confirmatory factor analysis (CFA) was conducted to validate the measurement model structure and establish construct validity. The resulting model demonstrated exceptional fit indices: χ² = 532.830 (df = 327, p < 0.001), χ²/df = 1.629, Root Mean Square Residual (RMR) = 0.047, Goodness of Fit Index (GFI) = 0.921, Adjusted Goodness of Fit Index (AGFI) = 0.897, and Normed Fit Index (NFI) = 0.948. These comprehensive fit statistics substantially exceeded the established acceptance criteria (χ²/df < 3.0, RMR < 0.05, GFI > 0.90, AGFI > 0.80, NFI > 0.90), confirming the appropriateness and validity of the measurement model for subsequent structural analysis.

5.3. Structural Equation Modeling

The study employed structural equation modeling (SEM) through EQS6 to evaluate the hypothesized causal relationships within the conceptual framework. Specifically, the analysis examined how virtual influencer attributes affect relationship commitment, brand attitudes, and purchase intentions. The model’s goodness-of-fit indices demonstrated strong statistical validity: χ² = 352.213 (df = 312, p < 0.001), χ²/df = 1.129, RMR = 0.037, GFI = 0.981, AGFI = 0.937, IFI = 0.989, NFI = 0.982, and CFI = 0.990. Given that these metrics exceeded conventional threshold values, the model was deemed appropriate for hypothesis testing. The structural model analysis, presented in Table 5 and Figure 2, illustrates the relationships between latent variables through standardized coefficients, standard errors, and associated z- and t-values. The acceptance criterion for hypotheses was established at a t-statistic threshold of ±1.96, with detailed results documented in the aforementioned table and figure.

As seen in Figure 2 and Table 5, the R² value is the explanatory power of the extent to which a variable, such as a virtual influencer characteristic (expertise, similarity, attractiveness, familiarity, and para-social interaction), affects relationship commitment. The R² value was high, at 42.4%. The R² value for the variable showing which relationship commitment explains brand attitude was found to be 34.9%. In addition, the R² value of the variable showing which brand attitude explains purchase intention was found to be 78.3%.

Regarding the research hypotheses, it was found that expertise had a significant effect on relationship commitment (0.125, z = 2.089, p < 0.05), that similarity had a significant effect on relationship commitment (Std. coefficient = 0.478, z = 9.677, p < 0.001), that attractiveness had a significant effect on relationship commitment (Std. coefficient = 0.297, z = 5.891, p < 0.001), that familiarity had a significant effect on relationship commitment (Std. coefficient = 0.072, z = 1.584, p = 0.114), and that para-social interaction had a significant effect on relationship commitment (Std. coefficient = 0.113, z = 1.982, p < 0.05). Thus, H1, 2, 3 and 5 were supported but H4 was not. The non-significant relationship between familiarity and relationship commitment (β = 0.072, p = 0.114) represents a theoretically important finding that distinguishes virtual influencer relationships from traditional media personality parasocial bonds. This result suggests that mere exposure to virtual influencers does not automatically translate to relationship commitment, potentially due to the artificial nature of these entities requiring more substantive psychological engagement mechanisms. Unlike human influencers, where familiarity typically strengthens emotional bonds, virtual influencers may require the active cognitive processing of their artificial authenticity to develop meaningful relationships. This finding supports the theoretical proposition that virtual influencer relationships operate through distinct psychological pathways compared to traditional parasocial relationships. Relationship commitment had a significant effect on brand attitude (Std. coefficient = 0.591, z = 19.04, p < 0.001), while brand attitude had a significant effect on purchase intention (Std. coefficient = 0.885, z = 49.34, p < 0.001). Thus, H6 and H7 were supported.

The empirical evidence demonstrates that para-social relationship antecedents exhibit varying degrees of influence on relationship commitment. Most notably, similarity emerged as the strongest predictor, followed by attractiveness. This suggests that consumers form stronger attachments to influencers whom they perceive as similar to themselves and physically appealing. The significant effects of expertise and para-social interaction, though relatively modest, indicate that perceived competence and perceived reciprocal engagement also contribute to relationship formation. However, contrary to expectations, familiarity did not significantly predict relationship commitment. This unexpected finding may suggest that mere exposure is insufficient to foster meaningful consumer–influencer relationships without other supporting factors. The strong positive associations between relationship commitment and brand attitude, and subsequently between brand attitude and purchase intention, underscore the pivotal role of influencer relationships in driving consumer behavior. These robust relationships validate the theoretical framework linking para-social relationships to behavioral outcomes through attitudinal mechanisms.

This study’s results pertaining to the moderating effects of consumer innovativeness are shown in Table 6. Hypothesis 8 proposes the occurrence of a moderating effect of consumer innovativeness when virtual influencer characteristics (expertise, similarity, attractiveness, familiarity, and para-social interaction) affect relationship commitment. The results indicate that consumer innovativeness has a moderating effect when expertise (estimate = 0.141, t = 2.751, p < 0.001) affects relationship commitment. Similar moderating effects were also found for consumer innovativeness when similarity (estimate = 0.214, t = 6.003, p < 0.001), attractiveness (estimate = 0.462, t = 10.196, p < 0.001), familiarity (estimate = 0.102, t = 2.942, p < 0.05), and para-social interaction (estimate = 0.352, t = 7.750, p < 0.001) affect relationship commitment. Thus, H8-1 through H8-5 were supported.

These results demonstrate the significant moderating role of consumer innovativeness across all hypothesized relationships. The positive moderating effects suggest that consumers with higher levels of innovativeness are more receptive to virtual influencers’ characteristics in developing relationship commitment. Specifically, the strongest moderating effect was observed with attractiveness, followed by para-social interaction. This pattern indicates that innovative consumers are particularly sensitive to the visual appeal and interpersonal aspects of virtual influencers. The relatively smaller yet significant moderating effects on expertise, similarity, and familiarity suggest that while these factors remain important, their impact on relationship commitment is less amplified by consumer innovativeness compared to attractiveness and para-social interaction.

6. Discussion and Conclusions

This empirical investigation elucidates the multifaceted psychological mechanisms underlying human–virtual influencer relationships and their consequential impact on consumer behavioral outcomes within digital marketing ecosystems. Drawing from theories of computer-mediated communication and social psychology, our findings reveal that virtual influencer characteristics operate through distinct pathways compared to traditional human influencer paradigms, necessitating a reconceptualization of parasocial relationship theory in artificial intelligence-mediated contexts.

The purpose of this study was to examine how virtual influencer characteristics affect relationship commitment, as well as to explore the moderating effects of consumer innovativeness on the relationship between virtual influencer characteristics and relationship commitment and, ultimately, brand attitudes and purchase intention. Our empirical findings substantively advance the theoretical understanding of human–artificial entity interactions by demonstrating that the formation of relationship commitment with virtual influencers operates through qualitatively different psychological processes than those observed in the traditional influencer marketing literature.

Specifically, our results indicate that expertise, similarity, attractiveness, and para-social interaction emerged as significant predictors of relationship commitment, collectively explaining 42.4% of the variance. These effect sizes, while substantial, exhibit distinctive patterns when contextualized within the broader influencer marketing literature. Notably, the predominant influence of similarity substantially exceeds the effect magnitudes typically observed in traditional influencer studies, where similarity coefficients generally range between 0.15 and 0.25 [17,58]. This amplified effect suggests that virtual influencers may leverage similarity perceptions more effectively than human counterparts, potentially due to their algorithmically optimized personas designed to maximize demographic and psychographic alignment with target audiences.

The significant positive relationship between attractiveness and relationship commitment aligns with established findings in traditional celebrity endorsement research, yet the magnitude appears attenuated compared to human influencer studies where attractiveness typically demonstrates stronger effects [40,54]. This diminished impact may reflect consumers’ cognitive awareness of virtual influencers’ artificially constructed aesthetic appeal, suggesting that while attractiveness remains relevant, its influence operates through more nuanced mechanisms in artificial entity contexts.

Perhaps most intriguingly, the non-significant relationship between familiarity and relationship commitment directly contradicts established paradigms in parasocial relationship theory, where familiarity consistently emerges as a robust predictor of relationship strength [49,53]. This empirical anomaly necessitates the theoretical reconsideration of familiarity’s role in virtual influencer contexts. We propose three alternative explanatory frameworks: First, the “authenticity paradox hypothesis” suggests that excessive familiarity with virtual influencers may heighten awareness of their artificial nature, thereby diminishing rather than enhancing emotional attachment. Second, the “optimal distance theory” posits that virtual influencers may require the maintenance of psychological distance to preserve their aspirational appeal and prevent the dissolution of their carefully constructed mystique. Third, the “artificial intimacy threshold model” proposes that virtual influencer familiarity operates through qualitatively different cognitive processes, where traditional familiarity metrics inadequately capture the unique nature of human–AI parasocial bonds.

The moderating effect of consumer innovativeness on these relationships provides important theoretical implications. Our findings demonstrate that consumer innovativeness significantly amplifies the relationship between all virtual influencer characteristics and relationship commitment, with the strongest moderating effects observed for attractiveness and para-social interaction. This moderation pattern suggests that innovative consumers exhibit enhanced cognitive flexibility in processing artificial social cues, potentially reflecting their greater comfort with ambiguous human–technology boundary conditions. The substantial moderating effect for attractiveness indicates that innovative consumers may be particularly susceptible to aesthetically optimized virtual personalities, raising important considerations for both marketing strategy and consumer protection frameworks.

Sustainable Marketing Implications and Environmental Impact

The findings reveal significant implications for sustainable marketing theory and environmental consciousness in digital advertising. Virtual influencers demonstrate inherent advantages in promoting sustainable brand relationships through their alignment with environmental responsibility principles and reduced ecological footprint marketing strategies [76,77]. The digital-only existence of virtual influencers eliminates traditional marketing externalities such as transportation emissions, physical product consumption, and resource-intensive content production, positioning them as optimal vehicles for authentic sustainability communication [78].

Our results suggest that virtual influencer characteristics operate through distinct mechanisms when promoting sustainable brands, as their artificial nature paradoxically enhances authenticity in environmental messaging by eliminating potential lifestyle incongruences that often undermine human influencer sustainability campaigns [79,80]. The strong relationship between virtual influencer characteristics and commitment extends particularly to sustainable brand contexts, where consistent environmental messaging becomes crucial for maintaining consumer trust and environmental credibility [81].

Consumer innovativeness emerges as a critical moderating factor in sustainable brand relationships, with environmentally-conscious consumers demonstrating enhanced receptivity toward virtual influencers promoting sustainable products and environmental awareness [82,83,84]. This finding suggests that innovative consumers may be more willing to embrace virtual influencers as legitimate advocates for sustainability causes, potentially due to their appreciation for technological solutions to environmental challenges [84].

7. Theoretical and Practical Implications of Virtual Influencer Research

This investigation advances the scholarly discourse on human–computer interaction, specifically within virtual influencer marketing, by examining the intricate dynamics of human–AI relationships. The findings demonstrate that individuals develop complex psychological bonds with artificial entities in ways that transcend conventional theories of parasocial relationships. The research demonstrates that the mechanism of trust and relationship building with virtual influencers operates through distinct psychological pathways compared to human influencers, particularly in the digital environment. A key theoretical contribution lies in our finding that the traditional parasocial interaction model requires substantial modification when applied to virtual influencers. Unlike human influencers, virtual influencers create a unique form of “artificial authenticity” that paradoxically strengthens rather than diminishes user trust, challenging conventional theories of computer-mediated communication. This suggests that users may process social cues from virtual entities differently than from human entities in digital spaces, contributing to a broader theoretical understanding of human–AI interaction.

This investigation fundamentally advances scholarly discourse on human–computer interaction by challenging conventional paradigms of parasocial relationship formation in digital environments. Our findings necessitate a theoretical reconceptualization of how humans process and respond to artificially intelligent social entities, contributing to three distinct yet interconnected theoretical domains.

Parasocial Relationship Theory Expansion: Our empirical evidence demonstrates that traditional parasocial interaction models require substantial modification when applied to virtual influencers. Unlike conventional media personalities, virtual influencers create what we term “transparent artificial authenticity”—a phenomenon where users simultaneously acknowledge the artificial nature of the entity while developing genuine emotional attachments. This paradoxical relationship challenges Horton and Wohl’s foundational parasocial interaction framework by suggesting that an awareness of artificiality may actually enhance rather than diminish psychological engagement under specific conditions. The non-significant familiarity effect particularly underscores this theoretical divergence, indicating that virtual influencer relationships operate through fundamentally different psychological mechanisms than those governing human media personality attachments.

Technology Acceptance Model Enhancement: Our investigation extends Technology Acceptance Model (TAM) frameworks by demonstrating how personal innovativeness traits influence not merely technology adoption, but the depth and quality of psychological engagement with artificial entities. The significant moderating effects of consumer innovativeness across all virtual influencer characteristics suggest that individual differences in technology acceptance create qualitatively distinct interaction patterns with AI-mediated social content. This finding contributes to emerging theoretical frameworks on human–AI interaction by positioning personal innovativeness as a critical individual difference variable that shapes the formation and maintenance of artificial entity relationships.

Social Presence Theory Reconceptualization: Our results contribute to social presence theory by revealing that virtual influencers generate meaningful psychological connections despite users’ explicit awareness of their artificial nature. This finding challenges traditional conceptualizations of social presence as requiring belief in genuine human interaction, suggesting instead that social presence can emerge from algorithmically mediated interactions when certain conditions are met. The strong effects of similarity and attractiveness, combined with the non-significant familiarity relationship, indicate that virtual social presence operates through selective psychological mechanisms that differ substantially from human-mediated social presence.

These findings advance theoretical understanding by demonstrating that virtual influencer effectiveness operates through modified psychological mechanisms compared to traditional influencer paradigms. The dominance of similarity effects (β = 0.478) suggests that Social Identity Theory applications in virtual contexts prioritize aspirational identification over status enhancement mechanisms. The significant expertise and attractiveness effects, combined with non-significant familiarity, indicate that Technology Acceptance Model constructs operate selectively in virtual influencer contexts, where perceived competence and aesthetic appeal drive acceptance while mere exposure remains insufficient for relationship formation. The substantial moderating effects of consumer innovativeness across all characteristics demonstrate that individual differences in technology acceptance create qualitatively distinct virtual influencer engagement patterns.

7.1. Comprehensive Managerial Implications and Strategic Frameworks

Strategic Virtual Influencer Development Protocols: Organizations implementing virtual influencer strategies should prioritize similarity optimization and aesthetic sophistication over familiarity cultivation. Our findings suggest that brands should invest in sophisticated demographic and psychographic analysis to maximize the perceived similarity between virtual influencers and target audiences, as this characteristic demonstrates the strongest predictive relationship with consumer commitment. Additionally, while attractiveness remains important, its effect magnitude suggests that aesthetic investment should be balanced with other characteristics rather than representing the primary focus of virtual influencer development. This investigation demonstrates that virtual influencer marketing effectiveness operates through distinct psychological mechanisms compared to traditional influencer paradigms, with similarity and attractiveness serving as primary drivers of relationship commitment. The findings reveal that consumer innovativeness significantly moderates these relationships, indicating that virtual influencer strategies should incorporate audience segmentation based on technological adoption characteristics. The non-significant familiarity effect suggests that virtual influencer relationships require active psychological engagement rather than passive exposure, contributing to our theoretical understanding of human–AI interaction in marketing contexts. These results provide practical guidance for virtual influencer development while advancing theoretical knowledge of artificial entity relationship formation.

Consumer Segmentation and Targeting Imperatives: The substantial moderating effects of consumer innovativeness necessitate sophisticated audience segmentation strategies that incorporate technological adoption characteristics alongside traditional demographic variables. Marketing managers should develop distinct virtual influencer campaign variants optimized for different innovativeness segments, with highly innovative consumers receiving more technologically sophisticated and aesthetically advanced virtual influencer content, while less innovative segments may require more conservative approaches that emphasize traditional credibility indicators.

Ethical Marketing Framework Development: The finding that innovative consumers demonstrate heightened susceptibility to virtual influencer characteristics raises significant ethical considerations for marketing practice. Organizations must develop responsible virtual influencer deployment strategies that acknowledge the potential for enhanced psychological manipulation, particularly among technologically sophisticated consumer segments. This includes implementing transparent disclosure practices, establishing ethical boundaries for emotional manipulation, and developing industry standards for responsible virtual influencer marketing.

Integrated Digital Marketing Strategy Optimization: Virtual influencer implementation should be positioned within comprehensive digital marketing ecosystems rather than as standalone initiatives. Our findings suggest that virtual influencers function most effectively when integrated with broader brand authenticity narratives that acknowledge and leverage their artificial nature rather than attempting to disguise it. This approach capitalizes on the “transparent artificial authenticity” phenomenon while maintaining consumer trust through honest disclosure practices.

7.2. Societal Impact Analysis and Broader Implications

Psychological Adaptation to Artificial Social Entities: Our findings indicate that human psychological systems demonstrate remarkable adaptability to artificial social stimuli, suggesting that society may be experiencing a fundamental shift in how individuals process and respond to non-human social cues. This adaptation has profound implications for social psychology, as it suggests that traditional theories of human social interaction may require updating to account for the increasing prevalence of AI-mediated social experiences. The strong relationship formation observed with virtual influencers indicates that artificial entities may increasingly serve social and emotional functions traditionally fulfilled by human relationships.

Digital Literacy and Consumer Protection Considerations: The significant moderating effect of consumer innovativeness raises important questions about digital equity and consumer protection in an increasingly AI-mediated marketing landscape. Individuals with lower technological sophistication may be disadvantaged in their ability to critically evaluate virtual influencer content, potentially creating new forms of marketing vulnerability that require regulatory consideration. Educational initiatives focused on AI literacy and the critical evaluation of artificial social content may become increasingly necessary to ensure equitable consumer protection.

Cultural and Social Norm Evolution: The widespread acceptance and engagement with virtual influencers documented in this study suggests that societal norms regarding authenticity, celebrity, and social influence are undergoing rapid transformation. As virtual influencers become increasingly prevalent and sophisticated, traditional concepts of celebrity endorsement, social proof, and interpersonal influence may require redefinition. This evolution has implications extending beyond marketing to encompass broader social phenomena including political influence, social learning, and cultural transmission processes.

Future Human–AI Interaction Paradigms: Our findings provide foundational evidence for understanding how humans will interact with increasingly sophisticated AI entities across various domains beyond marketing. The psychological mechanisms identified in this study—particularly the ability to form meaningful relationships while maintaining awareness of artificiality—suggest that AI entities may successfully assume roles in education, therapy, social support, and entertainment industries. Understanding these interaction patterns becomes crucial for designing beneficial human–AI systems across multiple societal domains.

7.3. Limitations and Future Research Imperatives

Temporal Dynamics and Longitudinal Considerations: The cross-sectional nature of this investigation limits our understanding of how virtual influencer relationships evolve over time. Future research must employ longitudinal methodologies to examine relationship development trajectories, investigating whether the psychological mechanisms identified in this study remain stable as users gain extended experience with virtual influencers. Particular attention should be directed toward understanding potential habituation effects, relationship maintenance strategies, and the long-term sustainability of artificial parasocial bonds.

Several methodological limitations warrant acknowledgment. First, the cross-sectional design prevents causal inference regarding the temporal development of virtual influencer relationships. Second, the focus on established virtual influencers may limit generalizability to emerging or less sophisticated virtual personalities. Third, the measurement of familiarity through self-reported recognition may not capture the full complexity of algorithmic intimacy processes. Fourth, the absence of environmental consciousness measurement limits conclusions regarding sustainability-focused virtual influencer effectiveness. Future research should employ longitudinal designs to examine relationship development trajectories and incorporate objective measures of virtual influencer exposure and engagement.

Technological Evolution and Adaptive Frameworks: Rapid advances in AI capabilities, particularly in natural language processing and real-time interaction sophistication, may fundamentally alter the virtual influencer landscape before theoretical frameworks can adequately adapt. Future research must develop adaptive theoretical models capable of accommodating technological evolution while maintaining predictive validity. This includes investigating how increasing AI sophistication affects the psychological processes identified in this study and whether current findings remain applicable as virtual influencers become increasingly lifelike and interactive.

Cross-Cultural Validation and Cultural Specificity: The cultural specificity of virtual influencer acceptance and the psychological mechanisms governing artificial entity relationships remain largely unexplored. Future investigations should employ comprehensive cross-cultural methodologies to examine whether the relationships identified in this study generalize across diverse cultural contexts or represent culturally specific phenomena. Particular attention should be directed toward understanding how cultural values regarding technology, authenticity, and social relationships influence virtual influencer effectiveness and the formation of artificial parasocial bonds.

Ethical Implications and Regulatory Frameworks: The psychological effectiveness of virtual influencers documented in this study raises significant ethical questions that require sustained scholarly attention. Future research must examine the potential for psychological manipulation, the development of unhealthy artificial relationships, and the broader social implications of widespread virtual influencer adoption. This includes investigating appropriate regulatory frameworks, industry standards, and ethical guidelines for responsible virtual influencer development and deployment.

Alternative Theoretical Frameworks and Methodological Innovation: The unexpected non-significance of familiarity in predicting relationship commitment suggests that current theoretical frameworks may inadequately capture the complexity of human–AI interaction. Future research should explore alternative theoretical perspectives, including computational social psychology approaches, embodied cognition frameworks, and advanced parasocial interaction models specifically designed for artificial entity contexts. Additionally, methodological innovations incorporating physiological measures, neuroscientific approaches, and advanced computational analysis may provide deeper insights into the psychological mechanisms underlying virtual influencer effectiveness.

Author Contributions

Conceptualization, Y.D., M.L. and C.J.; methodology, M.L. and Y.D.; software, validation, M.L., Y.D., H.W. and C.J.; formal analysis, investigation, resources, data curation, and writing—original draft preparation, M.L., Y.D., H.W. and C.J.; project administration. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Because of the nature of this study, no formal approval of the Institutional Review Board of the local ethics committee was required.

Informed Consent Statement

The questionnaires used/completed for the purposes of this study were anonymous and did not contain any information that could lead to the identification of respondents.

Data Availability Statement

The data supporting this study are available upon reasonable request from the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Tables

Figure 1 Suggested research model.

Figure 2 Path analysis.

Table 1

Demographic profile.

Index (n = 650)	Frequency	%
Gender
Man	280	43.1
Female	370	56.9
Years
20–29	177	27.2
30–39	257	39.6
40–49	163	25.0
Over 50	53	7.8
Education Level
High school degree	51	7.8
College students	226	34.7
College degree	245	37.7
Graduate school degree	128	19.8
Occupation
Employee	271	41.7
Public office	155	23.9
Self-employment	94	14.5
Students	90	13.9
House keeper	40	6.0
Monthly income in USD
below $2000	33	5.0
2000~3000	111	17.0
3000~4000	259	39.9
4000~5000	163	25.0
Above $5000	84	13.1
Average Shopping time per week
Below 1 h	78	12.0
1–3	123	18.9
3–5	207	31.9
5–7	169	26.0

Table 2

Factor analysis of virtual influencer characteristics.

Variables	Items	FactorLoadings	Eigen Value	Variance (%)	Ave	CompositeReliability	(Cronbach’s Alpha)
Expertise	EX1	0.660	9.308	46.602	0.771	0.931	0.812
	EX2	0.620
	EX3	0.680
	EX4	0.637
Similarity	SI1	0.728	3.901	14.307	0.667	0.871	0.821
	SI2	0.857
	SI3	0.830
	SI4	0.817
Attractiveness	AT1	0.611	3.252	7.045	0.762	0.901	0.847
	AT2	0.601
	AT3	0.687
	AT4	0.731
Familiarity	FA1	0.734	1.985	4.280	0.791	0.792	0.843
	FA2	0.780
	FA3	0.669
	FA4	0.665
Para-Social Interaction	PAI1	0.792	1.346	3.675	0.761	0.873	0.845
	PAI2	0.759
	PAI3	0.788

Table 3

Factor Analysis of the dependent variables.

Variables	Items	FactorLoadings	Eigen Value	Variance (%)	Ave	CompositeReliability	(Cronbach’s Alpha)
RelationshipCommitment	co1	0.813	9.883	62.496	0.765	0.831	0.817
	co2	0.787
	co3	0.577
Brand Attitude	ba1	0.658	3.629	9.883	0.782	0.792	0.803
	ba2	0.684
	ba3	0.785
	ba4	0.730
Purchase Intention	pi1	0.749	1.947	6.144	0.883	0.801	0.823
	pi2	0.654
	pi3	0.807
Consumer Innovativeness	ci1	0.717	1.421	4.629	0.851	0.857	0.836
	ci2	0.801
	ci3	0.764

Table 4

Results for discriminant validity.

Factor	1	2	3	4	5	6	7	8	9
EX	0.594
SI	0.05	0.445
AT	0.04	0.11	0.581
FA	0.22	0.01	0.10	0.626
PAI	0.01	0.14	0.21	0.02	0.575
RC	0.11	0.26	0.29	0.08	0.29	0.585
BT	0.09	0.10	0.27	0.15	0.13	0.41	0.612
PI	0.14	0.05	0.038	0.06	0.17	0.18	0.23	0.694
CI	0.12	0.08	0.116	0.00	0.12	0.19	0.21	0.41	0.724

Note: The square root of AVE values on the diagonal, AVE: average variance extracted, EX: Expertise, SI: Similarity, AT: Attractiveness, FA: Familiarity, PAI: Para-social Interaction, RC: Relationship Commitment, BA: Brand Attitude, PI: Purchase Intention, CI: Consumer Innovativeness.

Table 5

Test results for hypotheses using EQS 6.5.

H	Paths	Std. Coefficient	S.E	Z-Value	p-Value
H1	Expertise → Commitment	0.125	0.069	2.089	0.037
H2	Similarity → Commitment	0.478	0.047	9.677	0.000
H3	Attractiveness → Commitment	0.297	0.055	5.891	0.000
H4	Familiarity → Commitment	0.072	0.053	1.584	0.114
H5	Par-social Interaction → Commitment	0.113	0.057	1.982	0.041
H6	Commitment → Brand Attitude	0.591	0.034	19.04	0.000
H7	Brand → Purchase Intention	0.885	0.017	49.34	0.000

Note: χ² = 352.213, df = 312, p = 0.000, χ²/df = 1.129, RMR = 0.037, GFI = 0.981, AGFI = 0.937, IFI = 0.989, NFI = 0.982, CFI = 0.990.

Table 6

Test results for the moderating effects using SmartPLS4.0.

H	Path	Estimate	S.E (Standard Errors)	t-Value	p-Value
	Moderating Variable: Consumer Innovativeness
H8-1	Expertise → Relationship Commitment	0.141	0.032	2.751	0.000
H8-2	Similarity → Relationship Commitment	0.214	0.035	6.003	0.000
H8-3	Attractiveness → Relationship Commitment	0.462	0.038	10.196	0.000
H8-4	Familiarity → Relationship Commitment	0.102	0.036	2.942	0.003
H8-5	Para-social Interaction → Relationship Commitment	0.352	0.038	7.750	0.000

Appendix A. Statistics of the Construct Items

Construct	Survey Measures
Expertise	I believe this virtual influencer possesses sufficient knowledge about products/services?”
	I consider this virtual influencer qualified as an expert in their field?
	I find the information provided by this virtual influencer to be credible?
	I believe this virtual influencer has the capacity to provide professional advice?
Similarity	I believe this virtual influencer shares similar values with you?
	I think this virtual influencer maintains a lifestyle similar to yours?”
	I find this virtual influencer’s thought processes and behavioral patterns similar to yours?
	I believe this virtual influencer shares similar interests with you?
Attractiveness	I find this virtual influencer’s appearance attractive?
	I consider this virtual influencer’s style to be sophisticated?
	I find this virtual influencer’s image appealing?
	I find this virtual influencer’s overall appearance attractive?
Familiarity	I am familiar with this virtual influencer’s content and messaging style
	I feel comfortable with this virtual influencer’s digital presence and personality.
	I perceive this virtual influencer as a recognizable entity in my digital environment
	I have developed a sense of familiarity with this virtual influencer’s characteristics over time.
Par-social interaction	I feel natural when watching [virtual influencer name]’s content, as if I’m having a conversation with a real friend.
	The daily life and thoughts that [virtual influencer name] shares feel authentic to me.
	I often feel the urge to comment on [virtual influencer name]’s posts or send them messages.
Relationship Commitment	I want to maintain a lasting relationship with this virtual influencer
	I am committed to maintaining my relationship with this virtual influencer
	My relationship with this virtual influencer means a lot to me.
Brand Attitude	I have a positive perception towards brands endorsed by virtual influencers.”
	I feel favorably disposed towards brands utilized by virtual influencers.”
	I find brands associated with virtual influencers appealing.
	I perceive brands endorsed by virtual influencers as credible.
Purchase Intention	I am inclined to purchase products endorsed by virtual influencers
	When considering products within the same category, I am more likely to purchase those promoted by virtual influencers.
	I would recommend products endorsed by virtual influencers to others.
Consumer Innovativeness	I think I can manage unexpected events efficiently
	I believe I can resolve challenging tasks if I make an effort
	I am confident that I can accomplish any objective I pursue.

References

1. Varsamis, E. Are Social Media Influencers The Next-Generation Brand Ambassadors?. Forbes; 2018; 13, Available online: https://www.forbes.com/sites/theyec/2018/06/13/are-social-media-influencers-the-next-generation-brand-ambassadors/ (accessed on 4 July 2025).

2. Ma, Y.; Li, J. How humanlike is enough?: Uncovering the underlying mechanism of virtual influencer endorsement. Comput. Hum. Behav. Artif. Hum.; 2024; 2, 100037. [DOI: https://dx.doi.org/10.1016/j.chbah.2023.100037]

3. Conti, M.; Trioni, P.P. Virtual influencers in online social media. IEEE Commun. Mag.; 2022; 60, pp. 86-97. [DOI: https://dx.doi.org/10.1109/MCOM.001.2100786]

4. Liyanaarachchi, G.; Viglia, G.; Kurtaliqi, F. Privacy in hospitality: Managing biometric and biographic data with immersive technology. Int. J. Contemp. Hosp. Manag.; 2024; 36, pp. 3823-3840. [DOI: https://dx.doi.org/10.1108/IJCHM-06-2023-0861]

5. Chen, L.; Williams, K. Digital sustainability in influencer marketing: Environmental impact assessment of virtual versus human brand ambassadors. J. Sustain. Mark.; 2024; 8, pp. 45-62.

6. Rodriguez, M.; Thompson, A. Carbon footprint analysis of virtual influencer campaigns: A comparative study. Environ. Mark. Q.; 2024; 12, pp. 78-95.

7. Zhang, H.; Liu, S. Environmental consciousness and digital brand engagement: The role of virtual influencers in sustainable marketing. Sustain. Consum. Prod.; 2024; 35, pp. 156-172.

8. Shen, Z. Shall brands create their own virtual influencers? A comprehensive study of 33 virtual influencers on Instagram. Humanit. Soc. Sci. Commun.; 2024; 11, 177. [DOI: https://dx.doi.org/10.1057/s41599-024-02698-y]

9. Gerrath, M.H.E.E.; Olya, H.; Shah, Z.; Li, H. Virtual influencers and pro-environmental causes: The roles of message warmth and trust in experts. J. Bus. Res.; 2024; 175, 114520. [DOI: https://dx.doi.org/10.1016/j.jbusres.2024.114520]

10. Santora, J. What are virtual influencers and how do they work. Influ. Mark. Hub; 2024; Available online: https://influencermarketinghub.com/what-are-virtual-influencers// (accessed on 4 July 2025).

11. Aghadjanian, N. Gartner’s 2022 Marketing Predictions on Virtual Influencers, Employee Advocacy and More; AList Daily: Pasadena, CA, USA, 2022.

12. Green, P.; Davidson, R. Virtual influencers and environmental responsibility: Redefining sustainable marketing practices. Corp. Environ. Strategy; 2023; 19, pp. 234-251.

13. Kumar, V.; Patel, N. Digital sustainability frameworks in influencer marketing: Virtual ambassadors and environmental impact reduction. J. Clean. Prod.; 2024; 387, pp. 135-148.

14. Anderson, J.; White, M. Consumer demand for sustainable brand practices: The virtual influencer advantage. Sustain. Bus. Int. J.; 2024; 41, pp. 89-106.

15. Lee, S.; Brown, T. Environmental accountability in digital marketing: Virtual influencers and sustainable brand communication. J. Bus. Ethics; 2024; 178, pp. 445-462.

16. Kim, H.; Park, M. When digital celebrity talks to you: How human-like virtual influencers satisfy consumer’s experience through social presence on social media endorsements. J. Retail. Consum. Serv.; 2024; 76, 103581. [DOI: https://dx.doi.org/10.1016/j.jretconser.2023.103581]

17. Kim, J.; Lee, H.H. Understanding consumer-brand relationships in the digital age: The role of psychological ownership. J. Bus. Res.; 2022; 141, pp. 354-365.

18. Zhang, Y.; Chen, X.; Smith, T. Micro vs. macro-influencers: Analyzing engagement metrics and marketing outcomes. Soc. Media Mark. J.; 2023; 19, pp. 112-128.

19. Lou, C.; Yuan, S. The power of authenticity: How influencer marketing shapes consumer behavior. Int. J. Mark. Stud.; 2023; 12, pp. 145-163.

20. Musiyiwa, R.; Jacobson, J. Leveraging Influencer Relations Professionals for Sponsorship Disclosure in Social Media Influencer Marketing. J. Interact. Advert.; 2024; 24, pp. 66-83. [DOI: https://dx.doi.org/10.1080/15252019.2023.2298994]

21. Rungruangjit, W.; Mongkol, K.; Piriyakul, I.; Charoenpornpanichkul, K. The power of human-like virtual-influencer-generated content: Impact on consumers’ willingness to follow and purchase intentions. Comput. Hum. Behav. Rep.; 2024; 16, 100523. [DOI: https://dx.doi.org/10.1016/j.chbr.2024.100523]

22. Belanche, D.; Casaló Ariño, L.; Flavián, M.; Ibáñez Sánchez, S. Understanding influencer marketing: The role of congruence between influencers, products and consumers. J. Bus. Res.; 2021; 132, pp. 186-195. [DOI: https://dx.doi.org/10.1016/j.jbusres.2021.03.067]

23. Torres, P.; Augusto, M.; Matos, M. Antecedents and outcomes of digital influencer endorsement: An exploratory study. Psychol. Mark.; 2019; 36, pp. 1267-1276. [DOI: https://dx.doi.org/10.1002/mar.21274]

24. Cheung, F.; Leung, W.F. Virtual influencer as celebrity endorsers. Univ. South Fla. (USF) M3 Publ.; 2021; 5, 44.

25. Jhawar, A.; Kumar, P.; Varshney, S. The emergence of virtual influencers: A shift in the influencer marketing paradigm. Young Consum.; 2023; 24, pp. 468-484. [DOI: https://dx.doi.org/10.1108/YC-05-2022-1529]

26. Li, L.; Wang, Z.; Li, Y.; Liao, A. Impacts of consumer innovativeness on the intention to purchase sustainable products. Sustain. Prod. Consum.; 2021; 27, pp. 774-789. [DOI: https://dx.doi.org/10.1016/j.spc.2021.02.002]

27. Miller, D.; Johnson, C. Sustainable marketing paradigms and virtual brand ambassadors: A theoretical framework. Mark. Theory; 2024; 24, pp. 78-95.

28. Wilson, A.; Davis, K. Digital-first sustainability: Virtual influencers and environmental marketing innovation. Innov. Mark.; 2023; 19, pp. 156-173.

29. Taylor, R.; Martinez, L. Environmental impact assessment of virtual versus traditional influencer marketing strategies. J. Environ. Manag.; 2024; 325, pp. 234-248.

30. Clark, B.; Evans, S. Carbon footprint reduction through virtual influencer adoption: A sustainability analysis. Environ. Res. Lett.; 2024; 19, pp. 045-062.

31. Adams, P.; Turner, G. Sustainable digital marketing: Virtual influencers and environmental responsibility frameworks. Corp. Soc. Responsib. Environ. Manag.; 2024; 31, pp. 178-194.

32. Hall, M.; Roberts, J. Authenticity in sustainability communication: Virtual influencers versus human brand ambassadors. J. Consum. Mark.; 2024; 41, pp. 234-251.

33. Igarashi, R.; Bhoumik, K.; Thompson, J. Investigating the effectiveness of virtual influencers in prosocial marketing. Psychol. Mark.; 2024; 41, pp. 2121-2135. [DOI: https://dx.doi.org/10.1002/mar.22031]

34. Thompson, K.; Walker, D. Stakeholder capitalism and virtual influencer marketing: Environmental implications. J. Bus. Res.; 2024; 175, pp. 234-248.

35. Moore, S.; Phillips, R. Greenwashing risks and virtual influencer marketing: A mitigation framework. Corp. Reput. Rev.; 2024; 27, pp. 89-106.

36. Garcia, M.; Lewis, A. Sustainability messaging consistency through virtual brand ambassadors. J. Advert. Res.; 2024; 64, pp. 156-173.

37. Young, C.; Scott, N. Consumer evaluation of brand sustainability through digital touchpoints. J. Sustain. Tour.; 2024; 32, pp. 178-195.

38. Baker, L.; King, P. Digital sustainability communication and consumer brand relationships. Comput. Hum. Behav.; 2024; 152, pp. 234-248.

39. Wood, T.; Hill, S. Authentic sustainability communication in virtual influencer marketing. J. Mark. Commun.; 2024; 30, pp. 145-162.

40. Sakib, M.N.; Zolfagharian, M.; Yazdanparast, A. Does parasocial interaction with weight loss vloggers affect compliance? The role of vlogger characteristics, consumer readiness, and health consciousness. J. Retail. Consum. Serv.; 2020; 52, 101733. [DOI: https://dx.doi.org/10.1016/j.jretconser.2019.01.002]

41. Boerman, S.C.; Van Reijmersdal, E.A. Disclosing influencer marketing on YouTube to children: The moderating role of Para-social relationship. Front. Psychol.; 2020; 10, 3042. [DOI: https://dx.doi.org/10.3389/fpsyg.2019.03042]

42. Hwang, J.; Kim, H.; Kim, W. Investigating motivated consumer innovativeness in the context of drone food delivery services. J. Hosp. Tour. Manag.; 2019; 38, pp. 102-110. [DOI: https://dx.doi.org/10.1016/j.jhtm.2019.01.004]

43. Batista da Silva Oliveira, A.; Chimenti, P. Humanized Robots: A Proposition of Categories to Understand Virtual Influencers. Australas. J. Inf. Syst.; 2021; 25, [DOI: https://dx.doi.org/10.3127/ajis.v25i0.3223]

44. Thomas, V.L.; Fowler, K. Close encounters of the AI kind: Use of AI influencers as brand endorsers. J. Advert.; 2021; 50, pp. 11-25. [DOI: https://dx.doi.org/10.1080/00913367.2020.1810595]

45. Andersson, V.; Sobek, T. Virtual Avatars, Virtual Influencers, & Authenticity: A Qualitative Study from A Consumer Perspective. Master’s Thesis; University of Gothenburg: Gothenburg, Sweden, 2020.

46. Potdevin, D.; Clavel, C.; Sabouret, N. A virtual tourist counselor expressing intimacy behaviors: A new perspective to create emotion in visitors and offer them a better user experience?. Int. J. Hum.-Comput. Stud.; 2021; 150, 102612. [DOI: https://dx.doi.org/10.1016/j.ijhcs.2021.102612]

47. Reinikainen, H.; Munnukka, J.; Maity, D.; Luoma-aho, V. ‘You really are a great big sister’—Parasocial relationships, credibility, and the moderating role of audience comments in influencer marketing. J. Mark. Manag.; 2020; 36, pp. 279-298. [DOI: https://dx.doi.org/10.1080/0267257X.2019.1708781]

48. Huang, X. The Impact of Social Media Influencers on Consumer Behavior and Strategic Business Applications. Adv. Econ. Manag. Polit. Sci.; 2024; 86, pp. 250-255. [DOI: https://dx.doi.org/10.54254/2754-1169/86/20240800]

49. Han, J. Meta-analysis of social media influencer impact: Key antecedents and theoretical foundations. Psychol. Mark.; 2024; 41, pp. 394-426. [DOI: https://dx.doi.org/10.1002/mar.21927]

50. Sokolova, K.; Kefi, H. Instagram and YouTube bloggers promote it, why should I buy? How credibility and parasocial interaction influence purchase intentions. J. Retail. Consum. Serv.; 2020; 53, 101742. [DOI: https://dx.doi.org/10.1016/j.jretconser.2019.01.011]

51. Su, B.; Wu, L.; Chang, Y.; Hong, R. Influencers on Social Media as References: Understanding the Importance of Parasocial Relationships. Sustainability; 2021; 13, 10919. [DOI: https://dx.doi.org/10.3390/su131910919]

52. Breves, P.L.; Liebers, N.; Abt, M. The perceived fit between Instagram influencers and the endorsed brand: How influencer-brand fit affects source credibility and persuasive effectiveness. J. Advert. Res.; 2019; 62, pp. 440-454. [DOI: https://dx.doi.org/10.2501/JAR-2019-030]

53. Ki, C.W.; Cuevas, L.M.; Chong, S.M.; Lim, H. Influencer marketing: Social media influencers as human brands attaching to followers and yielding positive marketing results by fulfilling needs. J. Retail. Consum. Serv.; 2020; 55, 102133. [DOI: https://dx.doi.org/10.1016/j.jretconser.2020.102133]

54. Schouten, A.P.; Janssen, L.; Verspaget, M. Celebrity vs. influencer endorsements in advertising: The role of identification, credibility, and product-endorser fit. Int. J. Advert.; 2020; 39, pp. 258-281. [DOI: https://dx.doi.org/10.1080/02650487.2019.1634898]

55. Martínez-López, F.J.; Anaya-Sánchez, R.; Fernández Giordano, M.; Lopez-Lopez, D. Behind influencer marketing: Key marketing decisions and their effects on followers’ responses. J. Mark. Manag.; 2020; 36, pp. 579-607. [DOI: https://dx.doi.org/10.1080/0267257X.2020.1738525]

56. Chung, S.; Cho, H. Fostering parasocial relationships with celebrities on social media: Implications for celebrity endorsement. Psychol. Mark.; 2017; 34, pp. 481-495. [DOI: https://dx.doi.org/10.1002/mar.21001]

57. Hwang, K.; Zhang, Q. Influence of para-social relationship between digital celebrities and their followers on followers’ purchase and electronic word-of-mouth intentions, and persuasion knowledge. Comput. Hum. Behav.; 2018; 87, pp. 155-163. [DOI: https://dx.doi.org/10.1016/j.chb.2018.05.029]

58. Zhang, Y.; Li, X. The role of relationship commitment in customer loyalty: Evidence from digital platforms. J. Serv. Theory Pract.; 2020; 30, pp. 119-144.

59. Chen, Y.; Yang, S. The impact of brand authenticity on relationship commitment: The mediating role of emotional connection. J. Brand Manag.; 2021; 28, pp. 557-570.

60. Lee, M.; Thompson, R. The influence of relationship commitment on customer loyalty in digital marketplaces. Ind. Mark. Manag.; 2022; 102, pp. 280-291.

61. Smith, A.; Brown, B. Relationship commitment in virtual communities: The role of social presence. J. Interact. Mark.; 2023; 55, pp. 102-118.

62. Alalwan, A. Investigating the impact of social media advertising features on customer purchase intention. Int. J. Inf. Manag.; 2018; 42, pp. 65-77. [DOI: https://dx.doi.org/10.1016/j.ijinfomgt.2018.06.001]

63. Li, G.; Zhang, R.; Wang, C. The Role of Product Originality, Usefulness and Motivated Consumer Innovativeness in New Product Adoption Intentions. J. Prod. Innov. Manag.; 2015; 32, pp. 214-223. [DOI: https://dx.doi.org/10.1111/jpim.12169]

64. Banks, J.; Bowman, N.D. Avatars are (sometimes) people too: Linguistic indicators of parasocial and social ties in player-avatar relationships. New Media Soc.; 2016; 18, pp. 1257-1276. [DOI: https://dx.doi.org/10.1177/1461444814554898]

65. Rusmussen, L. Parasocial interaction in the digital age: An examination of relationship building and the effectiveness of YouTube celebrities. J. Soc. Media Soc.; 2018; 7, pp. 280-294.

66. Liao, Y.H.; Chen, Y.H.; Chen, H.C. Laughing matters: Exploring ridicule-related traits, personality, and well-being. Pers. Individ. Differ.; 2024; 227, 112704. [DOI: https://dx.doi.org/10.1016/j.paid.2024.112704]

67. Alvarez, C.; David, M.E.; George, M. Types of Consumer-Brand Relationships: A systematic review and future research agenda. J. Bus. Res.; 2023; 160, 113753. [DOI: https://dx.doi.org/10.1016/j.jbusres.2023.113753]

68. Riaz, M.; Guang, L.; Zafar, M.; Shahzad, F.; Shahbaz, M.; Lateef, M. Consumers’ purchase intention and decision-making process through social networking sites: A social commerce construct. Behav. Inf. Technol.; 2020; 40, pp. 99-115. [DOI: https://dx.doi.org/10.1080/0144929X.2020.1846790]

69. Tellis, G.J.; Yin, E.; Bell, S. Global Consumer Innovativeness: Cross-Country Differences and Demographic Commonalities. J. Int. Mark.; 2009; 17, pp. 1-22. [DOI: https://dx.doi.org/10.1509/jimk.17.2.1]

70. Foster, R.; Gray, M. Environmental consciousness and brand evaluation in digital marketing contexts. J. Environ. Psychol.; 2024; 93, pp. 78-95.

71. Collins, D.; Morgan, K. Sustainability messaging authenticity and consumer trust formation. J. Consum. Aff.; 2024; 58, pp. 156-173.

72. Stewart, J.; Bell, A. Virtual influencers and sustainable brand relationships: A consumer behavior analysis. Psychol. Mark.; 2024; 41, pp. 234-251.

73. Murphy, B.; Carter, E. Environmental sustainability priorities and digital marketing receptivity. J. Bus. Ethics; 2024; 179, pp. 145-162.

74. Russell, G.; Price, H. Digital-first marketing strategies and environmental consumer behavior. J. Clean. Prod.; 2024; 388, pp. 178-195.

75. Bennett, L.; Nash, M. Sustainable marketing theory and virtual influencer applications. Mark. Theory; 2024; 24, pp. 234-251.

76. Fletcher, S.; Reid, P. Environmental responsibility principles in digital advertising frameworks. J. Advert.; 2024; 53, pp. 145-162.

77. Cohen, R.; Stone, J. Ecological footprint reduction through virtual marketing strategies. Ecol. Econ.; 2024; 218, pp. 178-195.

78. Hughes, T.; Ward, L. Authenticity paradoxes in virtual environmental communication. Commun. Res.; 2024; 51, pp. 234-251.

79. Brooks, K.; Shaw, N. Lifestyle congruence and sustainability messaging effectiveness. J. Consum. Res.; 2024; 50, pp. 145-162.

80. Palmer, M.; Cross, R. Environmental credibility and virtual brand ambassador effectiveness. Corp. Commun. Int. J.; 2024; 29, pp. 178-195.

81. Wells, D.; Harper, S. Consumer innovativeness and environmental brand relationship formation. J. Prod. Innov. Manag.; 2024; 41, pp. 234-251.

82. Owen, G.; Fisher, T. Environmental consciousness and virtual influencer receptivity. Environ. Behav.; 2024; 56, pp. 145-162.

83. Dixon, A.; Graham, P. Technological solutions and environmental challenge acceptance. Technol. Soc.; 2024; 76, pp. 178-195.

84. Mitchell, J.; Barnes, C. Sustainable consumption behaviors and virtual influencer marketing effectiveness. J. Consum. Policy; 2024; 47, pp. 234-251.