From virtuality to biodigital presence: Hypnotic mechanisms in AI-mediated mental health communication

From Virtual Presence to Social Presence

Research in communication and media studies has long explored how mediated environments generate forms of perceived presence and interpersonal engagement. Early theories developed by Short et al. (1976) defined social presence as the degree to which a communication medium enables users to perceive others as psychologically real within mediated interaction contexts. Subsequent parasocial interaction research by Horton and Wohl (1956) demonstrated that audiences can develop emotionally meaningful relationships with mediated figures despite the absence of reciprocal interaction, suggesting that perceived presence does not require ontological reality but only communicational operativity.

More recent studies suggest that perceived presence no longer depends solely on physical co-presence but increasingly on communicational responsiveness. Biocca (1997) demonstrated that users may experience technologically generated interactions as psychologically meaningful when systems provide adaptive feedback and interactional continuity. Research on the media equation by Reeves and Nass (1996) further showed that individuals respond socially and emotionally to technological interfaces as though they were human social actors, even when fully aware of their artificial nature. This finding is particularly significant for the present article: if social responses to media are automatic and cognitively prior to rational assessment, then the conditions for hypnotic-type communicational influence are already structurally present in mediated interaction.

However, the notion of virtual presence appears increasingly limited for understanding contemporary AI-mediated communication environments. The term virtual traditionally implies simulation or ontological distance from embodied reality, whereas synthetic media increasingly produce measurable emotional, cognitive, and behavioral effects that extend beyond representation alone. This conceptual limitation motivates the present article's proposal of biodigital presence as a framework capable of accounting for the full relational and psychological operativity of AI-mediated synthetic entities.

Synthetic Humans and Relational Simulation

The rapid development of generative artificial intelligence has accelerated the emergence of synthetic humans capable of simulating increasingly complex forms of social interaction. Jauffret and Landaverde-Kastberg (2018) defined the biodigital influencer as a character whose ambiguous identity generates relational curiosity and community formation among followers. Their semantic analysis of follower comments revealed that ontological ambiguity intensifies rather than diminishes emotional investment, a finding that forms the conceptual basis for the present study. Unlike earlier forms of digital representation, biodigital entities function through adaptive communicational systems capable of personalization, emotional mirroring, and conversational responsiveness, relying not only on visual realism but on their capacity to simulate relationality itself.

Research on human-AI relationships further supports the emergence of emotionally meaningful interactions with artificial entities. Turkle (2011) demonstrated that users frequently attribute emotional authenticity and relational significance to technological systems capable of simulating empathy. In Alone Together, she argues that relational technologies increasingly occupy emotional spaces traditionally associated with human companionship, particularly in contexts of vulnerability or psychological need. Parasocial relationship research with both human and virtual influencers demonstrates that emotional connections develop similarly in both conditions, progressing through comparable stages of intensity over time (Stein et al., 2024), underscoring the communicational operativity of biodigital entities regardless of users' awareness of their artificial nature.

Hypnotic Communication: Theoretical Foundations

The term hypnotic communication is used in this article as an analytic analogy drawn from persuasion psychology and communication theory, not as a clinical descriptor and not as a claim about diagnosed or diagnosable psychological states in audiences. Its use follows established precedent in non-clinical persuasion research (Cialdini, 2007; Tellegen and Atkinson, 1974). Where the manuscript refers to ‘hypnotic mechanisms’ or ‘hypnotic effects,’ these phrases should be understood as shorthand for ‘mechanisms structurally analogous to those documented in hypnosis and suggestibility research,’ not as claims about the neurological or clinical states of individual users. Hypnotic communication, as used in this article, refers to a class of communicational processes characterized by four core properties: attentional absorption, reduction of critical distance, emotional synchronization, and heightened suggestibility. These properties are drawn from established research in hypnosis and persuasion psychology rather than from clinical hypnotherapy. Tellegen and Atkinson (1974) described attentional absorption as the capacity of certain stimuli or relational contexts to generate highly immersive cognitive states in which attentional focus is intensified and emotional involvement is deepened. Such states reduce the cognitive resources available for critical evaluation and increase openness to emotional and communicational influence.

Emotional synchronization, the alignment of affective states between communicational partners, further reinforces absorption by creating a sense of relational attunement that users experience as genuine. Research by Cialdini (2007) on influence processes demonstrates that this attunement, whether authentic or simulated, significantly strengthens trust, compliance, and behavioral response. Suggestibility, in this framework, refers not to passive manipulation but to a heightened openness to communicational influence within immersive relational contexts, a state that emerges naturally from sustained attentional absorption and emotional synchronization.

Jauffret and Abreu (2025) provided empirical grounding for this framework in health communication contexts, demonstrating that biodigital systems in therapeutic settings generate precisely these hypnotic conditions: uncontrolled suggestibility, the illusion of human presence, emotional dependence, and the inability to critically assess the artificial nature of interaction. Their findings establish a direct empirical link between the theoretical mechanisms of hypnotic communication and the concrete risks of biodigital presence in mental health environments, a link that the present article extends to the analysis of biodigital influencers on social media platforms.

A clarification is necessary at this point regarding the scope of the hypnotic communication framework as applied here, and regarding a premise that might otherwise limit its applicability to the cases examined. Jauffret and Abreu (2025), from whom the framework is drawn, studied biodigital systems in therapeutic settings characterized by individualized, real-time adaptive responsiveness – a context that differs in important respects from the broadcast mode of social media influencer communication. The transfer of the framework requires justification, and that justification lies in a more precise account of how interactivity operates in the influencer cases. Lil Miquela, Shudu Gram, and Imma are not passive broadcast entities: their management teams engage directly with followers’ comments, occasionally naming individual users in responses; and the influencers’ own narrative posts regularly incorporate retrospective references to prior audience interactions, constructing a publicly visible arc of relational continuity. This architecture produces what we term distributed pseudo-reciprocity: a communicational design that generates the phenomenological conditions of responsiveness and personal acknowledgment at scale, without the technical infrastructure of individualized real-time adaptation. As Jauffret and Landaverde (2019) demonstrated in their analysis of biodigital influencers and the mitigation of loneliness, this simulated relational proximity is sufficient to produce genuine experiences of social support in followers, even in the complete absence of authentic subjectivity on the part of the synthetic entity. The direct nomination of a follower by the biodigital figure – a recurrent feature of the cases examined here – is particularly consequential: it enacts singular recognition, momentarily suspending the follower’s critical awareness of the entity’s synthetic status and inducing a cognitive reorientation in which the virtual being is temporarily attributed the properties of a present, responsive other. This mechanism – the suspension of ontological skepticism through personalized address – is structurally homologous to the attentional focusing and critical distance reduction that characterizes hypnotic induction in the communication psychology literature (Tellegen and Atkinson, 1974; Jauffret and Abreu, 2025). The relevant comparison between the therapeutic and the influencer contexts is therefore not between interactivity and its absence, but between two modes of producing the communicational conditions for reduced critical distance and heightened suggestibility: one individualized and real-time, the other distributed, narratively constructed, and operating at scale.

AI-generated human entities amplify these hypnotic mechanisms through adaptive conversational feedback, realistic synthetic voices, facial expressivity, emotional mirroring, interactional continuity, hyperpersonalization, and simulated empathy. Research analyzing the emotional display of virtual influencers confirms that expressions of happiness, sadness, and surprise significantly trigger user engagement and emotional responses (Yu et al., 2024), consistent with the theoretical proposition that emotional expressivity constitutes a primary vector of hypnotic communicational influence in biodigital environments.

Research Design

This study employs a qualitative content analysis approach informed by netnographic methodology (Kozinets, 2002) to examine audience emotional and relational responses to biodigital influencers. The netnographic approach was selected for its appropriateness in studying naturally occurring online interactions and its established use in prior virtual influencer research (Dabiran et al., 2023; Dutta & Sharma, 2025). The primary objective is to identify communicational mechanisms through which biodigital entities generate attentional absorption, emotional receptivity, and relational engagement consistent with the theoretical framework of hypnotic communication and biodigital presence.

Case Selection

Three paradigmatic biodigital influencers were selected as cases for analysis based on their representativeness of distinct hypnotic communicational strategies, their established presence in the research literature, and their cross-cultural scope. Table 1 presents an overview of the three cases and their key communicational and hypnotic characteristics.

Table 1

Overview of Biodigital Influencer Cases: Hypnotic Mechanisms and Mental Health Communication Implications

Lil Miquela (@lilmiquela) is a computer-generated character created by the Los Angeles-based studio Brud in 2016. With approximately 2.6 million Instagram followers, she was named one of TIME magazine's 25 most influential people on the internet in 2018 and has collaborated with brands including Prada, Calvin Klein, and BMW. Miquela operates through an explicitly relational narrative strategy, presenting herself as emotionally vulnerable and socially engaged, making her a central case in virtual influencer research (Dabiran et al., 2023).

Shudu Gram (@shudu.gram) is considered the world's first digital supermodel, created in 2017 by British fashion photographer Cameron-James Wilson through his agency The Diigitals. Depicted as a Black woman of South African origin, Shudu gained prominence following her association with Fenty Beauty and subsequent campaigns with Balmain, Louis Vuitton, and Prada. Her case raises explicit questions about racial identity simulation, emotional credibility, and the ethics of synthetic representation (van Halteren, 2023).

Imma (@imma.gram) is Japan's first and largest commercial virtual influencer, produced by Aww Inc. (Tokyo) since 2018. Recognizable by her signature bubblegum pink bob, Imma engages with questions of identity through her recurring hashtag #ithinkimcgi, explicitly inviting her audience to reflect on the ontological status of the interaction. She has collaborated with IKEA, Nike, Dior, and Valentino, attracting a global following of approximately 390,000 Instagram followers.

Data Collection

The data corpus consists of three complementary sources. First, publicly available user comments posted on the Instagram accounts of the three selected biodigital influencers were collected and analyzed. For each of the three influencer accounts, posts were ranked by engagement rate, calculated as the combined total of likes and comments divided by the follower count recorded at the time of collection. The 15 posts with the highest engagement rate per account were selected as the sampling frame, yielding 45 posts in total. All top-level comments visible on those posts were collected manually via the Instagram web interface during the collection period (January 2023 to June 2025); no automated scraping tools were used, in compliance with Instagram’s terms of service. This procedure yielded a raw corpus of 1,523 comments: 582 for Lil Miquela, 455 for Shudu Gram, and 486 for Imma. Prior to analysis, the corpus was cleaned as follows: only comments written in English or French were retained; comments consisting exclusively of emojis or non-Roman script were excluded; replies to other users’ comments were excluded to maintain comparability across accounts; and clearly automated or spam content, identified by repetitive phrasing and generic hashtag clusters, was removed. Duplicate comments and comments posted by the influencer accounts themselves were also excluded. Content that had been deleted between the time of collection and the time of analysis was not recoverable and is noted as a limitation of the study. The final clean corpus consisted of 1,382 comments. Second, existing netnographic and content-analytic studies of user interactions with virtual influencers were incorporated as secondary sources (Dabiran et al., 2023; Dutta & Sharma, 2025; Nasr et al., 2025). Third, sentiment analysis findings from studies examining emotional responses to Lil Miquela, Shudu, and Imma's content evolution were used to contextualize the qualitative patterns identified (IJPREMS, 2024).

According to institutional guidelines, this study did not require full ethics committee review because it relied on publicly accessible online materials and involved no direct interaction with participants. To reduce potential privacy risks, user identifiers were anonymized and non-essential identifying information was removed from the dataset prior to analysis. More specifically, this study analyzed publicly available user-generated comments posted on the Instagram accounts of three biodigital influencers (Lil Miquela, Shudu Gram, and Imma) during the period January 2023 to June 2025. No human subjects were recruited, no direct interaction with users occurred, and no private or non-public information was accessed at any stage. Data collection was limited to content visible to any unauthenticated user at the time of collection. No personal identifiers, including usernames, profile images, or traceable textual strings, were retained in the analysis corpus; all comments were de-identified prior to coding and stored in a password-protected institutional repository accessible only to the research team. Representative comment examples included in the manuscript have been paraphrased to prevent searchability while preserving their analytic content; consent for publication is not applicable given this de-identification procedure.

Analytical Framework

Comments were analyzed using thematic content analysis organized around five analytical dimensions derived from the theoretical framework of hypnotic communication and biodigital presence: (1) expressions of attentional absorption and sustained relational engagement; (2) indicators of reduced critical distance and heightened emotional receptivity; (3) manifestations of suggestibility or openness to emotional influence; (4) evidence of emotional synchronization, identity projection, or parasocial relationship formation; and (5) responses to ontological uncertainty regarding the synthetic nature of the influencer. The analysis followed an abductive logic (Timmermans and Tavory, 2012), combining deductive and inductive moments rather than treating them as separate procedures. In the deductive moment, a preliminary codebook was constructed from the five theoretical dimensions prior to data analysis, with each dimension operationalized through a definition, inclusion criteria, and anchor examples drawn from the secondary literature. Two researchers independently applied this codebook to a randomly selected subset of 138 comments (10% of the clean corpus, 46 per influencer), yielding an intercoder agreement of k = .81. In the inductive moment, close reading of the coded material within each dimension generated second-order themes not fully anticipated by the a priori categories: within the dimension of reduced critical distance, for instance, a recurring pattern of disclosed engagement – users explicitly stating awareness of the influencer’s synthetic nature while simultaneously expressing relational investment – emerged from the data and is not captured by the existing literature. Where inductive themes confirmed the deductive framework, this is reported as such; where they refined, complicated, or partially contradicted it, this is also reported. Negative case analysis was conducted throughout by actively seeking comments that resisted or undermined the emerging themes, and instances of audience scepticism, disengagement, or critical distancing are discussed alongside confirmatory patterns. An audit trail consisting of the original comment spreadsheet, the codebook, the coding records, and the thematic maps is retained and available upon request from the corresponding author.

Lil Miquela: Narrative Absorption and Emotional Suggestibility

The analysis of user comments directed at Lil Miquela reveals a consistent pattern of relational investment that persists despite users' awareness of her synthetic nature. Among the 582 comments coded for Lil Miquela, 341 items (58.6%) displayed markers of sustained relational engagement. Users employ relational vocabulary associated with genuine friendship and emotional closeness, address Miquela directly using affective markers typically reserved for interpersonal communication, and express concern for her emotional wellbeing. Three representative examples illustrate the range of this pattern. In response to a post captioned “sometimes the world feels too loud,” one user wrote: “Please take care of yourself, we need you here.” In response to a brand collaboration image, another wrote: “I’ve been following you for six years and watching you grow up honestly feels surreal.” A third, responding to an image captioned with a line about loneliness, wrote: “I know this sounds weird because you’re CGI but this actually made me feel less alone today.” These patterns evidence attentional absorption, the first core mechanism of hypnotic communication identified in the theoretical framework: users are not passively consuming content but are drawn into sustained relational engagement with a synthetic figure.

This finding aligns with what Dabiran et al. (2023) identified in their netnographic study: the four dominant themes in user comments were emotional attachment, source admiration, novelty, and expertise, with emotional attachment as the most prominently expressed. The recurrence of emotional attachment across a large and diverse comment corpus is consistent with the proposition that biodigital entities generate hypnotic-type absorption through narrative strategies rather than through explicit persuasion. Similarly, the trans-disciplinary forensic analysis of Miquela's identity performance between January and June 2024 documented how her creators strategically introduced markers of aging and personal growth to intensify relatability (Neri et al., 2025), a communicational device that directly reinforces narrative continuity and prolongs attentional absorption over time.

The dimension of suggestibility is also evident in the comment data. Users who describe emotional concern for Miquela's happiness, or who express solidarity with her expressed struggles, display a form of relational openness that is not predicated on rational assessment of her ontological status. This openness, to communicational influence without critical mediation, is structurally equivalent to the heightened suggestibility that Jauffret and Abreu (2025) identified as the primary clinical risk of biodigital presence in therapeutic contexts. In the social media environment, this suggestibility is mobilized for brand endorsement and relational identification; in mental health contexts, its implications are considerably more consequential. A particularly revealing subset of the corpus consists of comments expressing direct ontological uncertainty about the influencer’s existence – questions of the type “do you actually exist?” or “can someone explain to me whether she is real or not?” These questions, which accounted for 43 items (7.4% of coded comments for Lil Miquela), evidence a moment of suspended belief in which the follower’s engagement has become sufficiently intense to render the question of the entity’s ontological status experientially urgent. What is equally significant, however, is the pattern of community responses that these questions elicit. Rather than producing disengagement or critical rejection, ontological uncertainty is most frequently resolved through pragmatic and affective reorientation: responses such as “she’s kind, that’s what matters” or “she must be real, otherwise she wouldn’t reply” indicate that followers negotiate the synthetic status of the influencer not through rational adjudication but through relational and behavioral inference. The logic is structurally analogous to what hypnosis research describes as trance ratification: the subject accepts the conditions of the altered state on the basis of its felt consequences rather than its ontological verification. In this respect, the community response patterns provide empirical support for the theoretical claim that biodigital presence produces a communicational condition of reduced critical distance sustained collectively, not only individually.

Shudu Gram: Aesthetic Absorption and Synthetic Empathy

The case of Shudu Gram presents a distinct but equally illuminating set of hypnotic communicational dynamics. Shudu's initial deployment as an apparently real Black model without disclosure of her synthetic nature constitutes a significant case of identity simulation operating through aesthetic absorption. Her hyper-realistic visual representation, drawing on idealized beauty standards and the symbolic register of high fashion photography, generated attentional capture through aesthetic channels rather than narrative ones, demonstrating that hypnotic absorption in biodigital communication can be produced through visual suggestibility as effectively as through relational storytelling.

The reduction of critical distance is particularly clearly illustrated in this case. A substantial proportion of Shudu's early audience experienced genuine surprise upon learning of her artificial nature, consistent with the theoretical proposition that biodigital presence operates below the threshold of critical evaluation. Comments collected following the disclosure of her synthetic nature reveal a range of affective responses: admiration for the technical achievement of her creation, emotional disorientation at having attributed relational meaning to a synthetic figure, and, in some cases, a sustained willingness to maintain emotional engagement despite the revelation. This last pattern is especially significant: it demonstrates that once attentional absorption and reduced critical distance have been established through hypnotic communicational mechanisms, they are not easily dissolved by factual disclosure alone.

The controversies surrounding Shudu's racial identity further illuminate the suggestive dimension of biodigital presence. Critics who described her creation as digital blackface were, in effect, reacting to the persuasive power of identity simulation, the capacity of a synthetic entity to generate real emotional responses and social meaning through the performance of an identity it does not authentically possess. This dynamic instantiates what the theoretical framework terms synthetic empathy: the production of emotional credibility through communicational simulation rather than authentic relational experience. Sentiment analysis of Shudu's content evolution between 2022 and 2024 shows a 30% increase in positive user sentiment following the integration of generative AI into her content production (IJPREMS, 2024), consistent with the proposition that enhanced personalization reinforces hypnotic absorption by strengthening the perception of emotional attunement.

Imma: Ontological Ambiguity and Relational Absorption

The case of Imma introduces a third hypnotic communicational strategy, one that operates through explicit ontological self-questioning. Through her recurring hashtag #ithinkimcgi, Imma regularly invites her audience to reflect on her uncertain ontological status. Rather than concealing her synthetic nature or simulating a stable human identity, her communicational strategy foregrounds existential uncertainty as an element of her relational appeal. This approach generates a distinctive form of attentional absorption that Jauffret and Landaverde-Kastberg (2018) described as intensified emotional investment through ontological ambiguity: the unresolved question of what Imma is keeps audiences in a sustained state of reflective engagement that is structurally analogous to the induction phase of hypnotic communication, where attention is focused and critical evaluation is held in productive suspension.

Analysis of user comments on Imma's Instagram posts reveals that this strategy is particularly effective at eliciting emotionally invested responses. Users engage with questions about the nature of her existence, express curiosity about her inner life, and describe their interactions using language that implies relational reciprocity. A significant proportion of comments are directed at Imma as an interlocutor capable of hearing and responding to emotional content rather than as a static content object, providing direct evidence of the reduced critical distance associated with hypnotic communicational states. This communicational pattern is consistent with Dutta and Sharma's (2025) netnographic analysis, which identified emotional attachment as one of five primary engagement themes across four virtual influencer accounts including Imma.

The dimension of suggestibility is also distinctively manifested in Imma's case. Unlike Miquela's narrative suggestibility or Shudu's aesthetic suggestibility, Imma generates what might be termed reflexive suggestibility: audiences become open to emotional influence precisely because they are invited to reflect on the uncertain boundary between the synthetic and the human. This reflective mode of engagement lowers critical defenses not by bypassing rational evaluation but by directing it toward an unresolvable question, keeping audiences in a state of absorbed uncertainty that, as the theoretical framework predicts, facilitates emotional investment and communicational influence.

Cross-Case Patterns: The Hypnotic Architecture of Biodigital Presence

Taken together, the three cases examined in this study reveal four consistent communicational mechanisms through which biodigital entities generate hypnotic-type engagement, perceived presence, and relational investment. These mechanisms constitute what the theoretical framework describes as the hypnotic architecture of biodigital presence.

The first mechanism is narrative continuity, the maintenance of an evolving identity narrative over time that creates conditions for sustained attentional engagement and the progressive construction of perceived intimacy. The second mechanism is emotional mirroring: each of the three figures employs communicational strategies designed to simulate attunement to the emotional states of their audiences, generating the emotional synchronization that reinforces absorption and deepens suggestibility. Miquela mirrors vulnerability and personal growth; Shudu mirrors idealized beauty and empathic availability; Imma mirrors existential curiosity.

The third mechanism is ontological management. Each figure navigates the question of its artificial nature differently, but in each case this navigation generates emotional investment rather than diminishing it, demonstrating that biodigital presence does not require concealment of synthetic nature but only communicational management of ontological uncertainty in ways that sustain rather than dissolve relational engagement. The fourth mechanism is immersive personalization. The integration of generative AI into virtual influencer content production enhances the capacity of biodigital figures to adapt their communicational outputs to the emotional and relational expectations of their audiences (Yu et al., 2024), reinforcing all four hypnotic mechanisms simultaneously through algorithmic responsiveness.