IntroductionLarge language models (LLMs), such as GPTs, are increasingly being considered as tools for simulating human responses to communication. This paper investigates whether LLM-based AI agents can reproduce nuanced, identity-based responses to persuasive health messages, with a focus on political and religious identity cues in COVID-19 vaccine communication.MethodsAcross two studies, AI agents were assigned political or religious identities and exposed to COVID-19 vaccine messages that varied in the presence and congruence of identity-based cues. Study 1 examined whether politically identified AI agents responded differently to vaccine messages endorsed by politically congruent versus incongruent figures. Study 2 examined whether AI agents assigned varying levels of Christian religiosity responded differently to vaccine messages that included versus omitted religious identity cues.ResultsIn Study 1, AI agents' responses aligned with their assigned political affiliations, with greater persuasion occurring when vaccine messages were endorsed by politically congruent figures. In Study 2, AI agents identified as moderately or highly religious Christians responded more favorably to vaccine messages incorporating religious identity cues than to messages without such cues.DiscussionThe findings suggest that LLMs can simulate nuanced, human-like responses to identity-based health communication, highlighting their potential value for pretesting and refining public health messages. The results also reinforce the persuasive potential of identity-based cues in health communication. More broadly, the studies demonstrate the promise of LLM-based AI agents as a methodological tool for persuasion and health communication research while pointing to important directions for future research on the validity and appropriate use of simulated human responses.
Artificial Intelligence (AI) has the potential to transform advergames by enabling more adaptive, personalized, and data-driven interactions between brands and consumers. While previous research has established the effectiveness of advergames in generating consumer engagement and brand-related responses, less attention has been given to the distinction between user-directed personalization and algorithmic, AI-driven personalization in branded gaming environments. This study examines the intersection of personalization, interactivity, brand integration, and AI in contemporary advergames, with particular attention to consumer engagement, brand recall, and ethical considerations. The study employs a qualitative comparative case study approach. Four prominent branded gaming campaigns NIKELAND, Coca-Cola Zero Sugar Byte, Gucci Garden, and Wendy's Food Fight/Keeping Fortnite Fresh were examined using publicly available secondary sources, including company and platform materials, campaign reports, news and industry articles, performance statistics, and relevant academic literature. The cases were comparatively analyzed across four dimensions: game design and interactivity, brand integration and advertising strategy, consumer engagement and publicly reported performance, and ethical considerations related to personalization and data use. The analysis indicates that successful advergames create engagement primarily through the integration of branded content with interactive gameplay, participation, social interaction, and immersive experiences. The cases demonstrate different forms of user-directed customization, social participation, exclusivity, digital collectibles, and brand integration. These features can strengthen consumer involvement and brand associations, but the publicly available evidence does not establish that all four campaigns employed AI-driven personalization or algorithmic adaptation. The findings therefore distinguish between personalization as a design feature and AI-driven personalization based on real-time behavioral data and algorithmic adaptation. The analysis also highlights privacy, transparency, and consumer autonomy as important considerations when personalization relies on the collection and analysis of player data. The study contributes to the literature by conceptualizing AI-driven personalization in advergames as an intersection of personalization, interactivity, immersion, and algorithmic adaptation rather than treating all forms of customization as AI. The findings suggest that brands can strengthen advergame effectiveness by aligning gameplay mechanics with brand identity and providing meaningful opportunities for participation, while the responsible use of AI requires greater transparency regarding data collection, personalization mechanisms, and user control. The study also identifies the need for future empirical research examining proprietary AI systems, behavioral data practices, and the effects of algorithmic personalization on engagement, brand recall, trust, and consumer autonomy.
IntroductionA substantial computational literature reports that audience reception can be predicted from the language of film plot summaries. We tested that claim across two independent corpora that differ systematically in summary register and length.MethodsCorpus A comprises 8,551 platform summaries from The Movie Database (TMDb) across 46 original languages (median 44 words), whereas Corpus B comprises 3,802 encyclopaedic plot summaries from the CMU Movie Summary Corpus matched to TMDb ratings (median 672 words). An identical pipeline extracted features representing narrative accessibility, affective engagement, and cultural legibility. We assessed feature -rating associations, cross-market differences, and predictive performance relative to majority-class and bag-of-words baselines.ResultsAcross both corpora, the largest correlation between any linguistic feature and mean audience rating was r = 0.08, corresponding to less than 1% shared variance. Films succeeding across markets did not differ from films succeeding within their home market in cultural specificity (d = −0.05 and −0.08) or universal thematic content (d = 0.06 and −0.04). No engineered-feature model exceeded the majority-class baseline. The TF-IDF baseline outperformed the complete engineered feature set: 33.9% versus 28.2% accuracy in Corpus A and 43.3% versus 31.5% in Corpus B; in Corpus B, TF-IDF also produced R² = 0.201 versus 0.043 for the best engineered model. Cultural specificity varied across original-language groups in Corpus B (η² = 0.050) but predicted neither ratings nor cross-market success.DiscussionThe linguistic properties of plot summaries carry little recoverable information about audience reception. Because Corpus B produced well-distributed values for measures that were degenerate in Corpus A, the result is unlikely to reflect measurement failure alone. Range restriction and popularity confounding remain plausible partial explanations, and previously reported high predictive performance warrants re-examination against explicit baselines.
Generation Z in Greater Jakarta engages with thrifting as an everyday communicative practice through which ideas about sustainability are understood and expressed. This study examines this practice through Social Practice Theory, Sustainable Consumption Theory, Media Ecology Theory, and Attribution Theory. Using a qualitative interpretivist design and reflexive thematic analysis, the study draws on in depth interviews with seven purposively selected participants. Their accounts indicate that thrifting is associated with affordability, product quality, identity expression, environmental awareness, and engagement with digital media. Participants also recognized the environmental benefits of secondhand clothing while considering broader consequences of imported thrift products, including transportation emissions, implications for local textile industries, and global waste circulation. Instagram and TikTok emerged as communicative spaces where sustainable fashion became more visible through algorithmic exposure, peer interaction, and user generated content. Across these accounts, sustainability was not understood solely as an environmental concern. It was connected to economic value, cultural identity, environmental responsibility, and experiences with digital media in everyday thrifting practices. The study contributes to environmental communication scholarship by showing how meanings of sustainable fashion are shaped through the intersection of consumption, communication, and digital media in a Global South context.
PurposeGeneration Z cohort is often more susceptible to social media influencers, as they are more involved with and engaged in social media compared to other cohorts. This paper explores the effect of social media influencer marketing on young consumers and how it shapes their decision-making and purchase behaviour.Design/methodology/approachThe paper used SPAR-4 and the TCCM framework to systematically synthesise the literature from 60 studies focused on young consumer behaviour and influencer marketing. The study aims to identify influencer attributes that shape young consumers' trust, brand perception, brand attitude, brand image, and overall purchase behaviour, including their purchase intent and actual purchase decisions.FindingsIt offers a comprehensive understanding of the elements that impact youth trust, purchase behaviour and psychological barriers surrounding the influencer marketing phenomenon.Research limitations/implicationsWhile the study adopts a multi-disciplinary approach to review the existing literature and systematically synthesise and summarise the findings, the articles were retrieved from a selected database, and any study falling beyond those search criteria was excluded.Social implicationsThis paper draws attention to the role of social media influencer marketing on the holistic well-being of young adults.Practical ImplicationsThe study provides meaningful guidance for brand marketers on effectively selecting influencers and formulating marketing strategies for younger audiences.OriginalityThe paper proposes future propositions and a nomological framework to further inform research in this direction. This study elaborates on the current body of literature, helping brand marketers develop and adopt influencer-led strategies, particularly for younger audiences.
This article reviews the methodologies of two research approaches that share a common interest: Phonetic variation in (natural, spontaneous) speech. The first approach, Conversation Analysis/Interactional Linguistics, is strictly based on qualitative analyses, particularly sequence analysis. Studies within the second approach, ‘laboratory’ or ‘empirical phonology’, use quantitative methods. The article begins by contrasting the two approaches in terms of their typical methods. It argues that quantitative methods can and should also be applied within qualitatively oriented CA/IL. It first discusses the compatibility of interactional analysis and statistical evaluation. It then addresses an area that poses a challenge for both approaches in equal measure, namely the question of conceptualising phonetic reduction. It is argued that quantitative approaches can benefit from CA/IL in this regard.
This article examines Evi Stamou's Who Was Here? (2025) as a documentary that uses generative artificial intelligence to work with a fragile family archive and an inaccessible historical past. The film combines family photographs, oral memory, written prompts, ChatGPT responses, and AI-generated images to approach the silence surrounding the director's father and his recruitment during the Turkish invasion of Cyprus in 1974. Rather than treating generated images only through the risk of falsification, the article argues that Who Was Here? shifts attention toward custody: who can sustain a story when the archive is incomplete, politically muted, or difficult to transmit? Through an analysis of the interaction between the director and artificial intelligence, the generated reconstructions, and the distortions of the images, the article shows how Stamou turns generative tools into a counter-narrative device. The director preserves errors and introduces corrections in order to mark the distance between the plausibility produced by the models and the situated knowledge of the events transmitted to her. The article proposes the term algorithmic mourning to describe this incomplete attempt to keep present what has been lost by working with its remains through generative forms. Ultimately, Who Was Here? suggests that machines can assist remembrance, but cannot assume memory as responsibility.
This study examines Kazakhstani youth's reading and media-consumption shifts toward digital platforms and analyzes how these shifts shape the use and perceived availability of Kazakh-language content in contemporary digital communication. Drawing on an online survey of young people conducted between February and April 2024 (N = 1,570; aged 14–35) across all 20 administrative regions of Kazakhstan, the study examines language practices (language of thought and language of online communication), reading-format preferences, and content demand, and tests their relationship with the frequency of Kazakh-language media consumption and satisfaction with the quantity and quality of Kazakh-language digital content. The findings are interpreted as applying to the study sample and as a basis for future replication using probability-based samples. Descriptive statistics and inferential analyses indicate that the digital environment is now the dominant channel for youth information consumption, accompanied by reduced reliance on print and broadcast formats. Results show that language practices are strongly associated with Kazakh-language engagement: respondents operating primarily in Russian or English digital environments report substantially lower frequency of consuming Kazakh-language content than those who operate primarily in Kazakh, while demographic effects are more limited. The findings also suggest uneven satisfaction with Kazakh-language content availability, pointing to gaps in informational and educational/scientific materials. The study concludes that strengthening Kazakh-language presence in new media requires localization and content strategies that improve visibility and discoverability across platforms, expand high-demand knowledge content, and support language policy goals through digitally native distribution and community-oriented communication.
The emergence of historical channels mediated by artificial intelligence (AI) on social media platforms constitutes an emerging communicative phenomenon that the academic literature in visual communication has not yet addressed in a systematic way. Chloe VS History, a YouTube and TikTok channel, offers a paradigmatic case of this new digital narrative. Based on three consolidated theoretical frameworks: the grammar of visual design by Kress and van Leeuwen, the gatekeeping theory by Shoemaker and Vos, and the concept of epistemic injustice by Fricker, the convergence construct of visual hyperrealism is proposed to name the simultaneous intersection of three conditions: the production of images of technical fidelity indistinguishable from a real record, a distribution environment without institutional verification mechanisms, and a good part of audiences without critical tools to evaluate that content. It is proposed that this convergence could generate a form of new epistemic risk that none of the three frameworks, in isolation, manages to fully capture, and that would help explain how a historical inaccuracy could be perceived as true without encountering any institutional or cognitive barrier. The central problem is not the use of AI in itself, nor the story as content, nor entertainment as a format, but its specific articulation in an ecosystem where visual credibility operates independently of factual fidelity. Ethical implications are derived for the individual production of historical content with AI and for the role of distribution platforms, and the analytical limits of the case are discussed in order to open an empirical research agenda in this emerging field.
IntroductionOver-the-top (OTT) services must retain users and encourage favorable recommendations in an increasingly competitive digital entertainment market. This study examined how content perceptions and perceived ease of use influence brand trust and customer engagement, and how these constructs relate to OTT continuance intention and positive word-of-mouth intention among users in Chile.MethodsA cross-sectional survey was conducted in Chile between August and September 2024. Data from 282 OTT users were analyzed using partial least squares structural equation modeling (PLS-SEM), complemented by importance-performance map analysis (IPMA). Drawing on the Stimulus-Organism-Response (S–O–R) framework, content perceptions and perceived ease of use were conceptualized as service-related stimuli, brand trust and customer engagement as organism-level states, and OTT continuance intention and positive WOM intention as post-adoption responses.ResultsContent perceptions positively influenced brand trust and customer engagement, whereas perceived ease of use positively influenced brand trust but not customer engagement. Brand trust and customer engagement positively influenced OTT continuance intention and positive WOM intention, while OTT continuance intention also positively influenced positive WOM intention. IPMA identified customer engagement as the main improvement priority for both OTT continuance intention and positive WOM intention, combining relatively high importance with comparatively low performance.DiscussionThese results indicate that different service-related stimuli activate distinct organism-level mechanisms: content perceptions are associated with both confidence- and involvement-based pathways, whereas perceived ease of use is associated primarily with a trust-based pathway. Content perceptions emerge as the main upstream driver of favorable post-adoption outcomes, while the IPMA findings highlight customer engagement as a particularly actionable managerial priority for strengthening both retention and advocacy. These findings extend the S–O–R framework to OTT post-adoption behavior in Chile by showing that service-related stimuli and organism-level states play different roles in generating retention and advocacy.
Generative AI is accelerating audiovisual production in games and extended reality (XR), yet it also introduces a distinct socio-technical risk: creators can produce complex, plausible outputs faster than they can meaningfully verify, maintain, or author them. We term this phenomenon Digital Bravado - a misalignment between perceived competence and actual capability that arises when knowledge remains externalized in the model rather than internalized by the creator. While related to automation bias, AI overreliance, the Dunning-Kruger effect, and cognitive offloading, Digital Bravado is domain-specific: it is particularly consequential in interactive audiovisual production, where meaning is enacted through motion, timing, and feedback. Drawing on research-creation methods grounded in the authors' full development and dual-store release of the mobile interactive experience Sleep Train, this Conceptual Analysis proposes a four-layer competency model (AI Literacy → Production Infrastructure Literacy → Engine and Tool Integration Literacy → Critical Literacy) to help creators and educators use generative tools as collaborators without surrendering voice, intentionality, or care. Two practical artifacts, a model overview and a self-assessment rubric, are offered as propositional tools to support curriculum design, reflective practice, and further validation. The framework aims to make Digital Bravado visible and to provide a developmental pathway beyond it.
Communication failures remain a leading cause of sentinel events in healthcare settings, yet the mechanisms by which organizational communication hierarchies form, function, and fail remain undertheorized. Prior work reconceptualized the Balanced Scorecard (BSC) as a constitutive communication mechanism through Communication Constitutes Organization (CCO) four flows theory; however, a principled account of why such hierarchies erode over time was lacking. This paper proposes that three functional gradient axes of the human neocortex—sensory-association (SA), modulation-representation (MR), and visual-somatosensory (VS)—offer a disciplined analogy for the communication hierarchy of healthcare organizations: the SA axis maps onto vertical strategic integration in BSC implementation; the MR axis parallels the tension between organizational self-structuring and activity coordination in CCO theory; and the VS axis corresponds, more loosely, to interprofessional communication differentiation and integration. Drawing on the neuroscientific finding that gradient differentiation follows an inverted-U trajectory and declines through dedifferentiation, the concept of organizational dedifferentiation is introduced: progressive erosion of communication hierarchy in mature healthcare institutions, manifesting as SA collapse, MR rigidification, and VS homogenization. Unlike existing lifecycle theories, this framework offers a candidate mechanism, a tentative collapse sequence, and three diagnostic indicators—SA coherence index, MR balance ratio, and VS integration score—proposed as measurable in principle before overt communication failure occurs. Active constitutive BSC use is proposed as a functional organizational analogue of cognitive reserve—by analogy rather than biological equivalence—that may help sustain communication hierarchy against dedifferentiation. This perspective opens a new interdisciplinary research agenda at the intersection of organizational communication theory and cognitive neuroscience.
Science podcasts are increasingly prominent in public-facing science communication, but little is known about how listeners evaluate their credibility under everyday audio conditions. This exploratory qualitative study used a shared episode from a German public-service science podcast and three focus groups with ten adults. The transcripts were analyzed using deductive-inductive structuring qualitative content analysis. Participants’ accounts were organized around two functionally distinct but related dimensions. The affective-closeness dimension captured relational responses associated with voice and prosody, narrative self-disclosure, direct address and interaction, and familiarity through repeated listening. The cognitive-transparency dimension captured participants’ perceptions that communicative features supported access to evidence, epistemic status marking, and structural orientation. Participants often related approachable, relational communication to the expectation that the basis and status of scientific claims could be located and examined when needed. The findings do not establish that credibility consists of these dimensions or that a particular balance between them produces credibility. Rather, they support a context-bound, empirically grounded interpretative model of credibility evaluation in the shared-stimulus setting. The study further derives a preliminary, listener-informed checklist for reflective podcast production and publication. This heuristic requires validation across audiences, genres, topics, production settings, and cultural contexts.
IntroductionSustainability problems such as biodiversity loss are difficult to communicate due to their complex, interrelated systems, divergent perspectives and controversies, scientific uncertainties, and pressing political decisions.MethodsBased on qualitative interviews and document analysis, this study investigates how the public communication of wicked sustainability problems is realised in a natural history museum.ResultsIt finds that the museum staff employs the Dissemination, Dialogue, and Participation communication models to address the inherent characteristics of wicked sustainability problems, namely Complexity, Uncertainty, Controversy, and Urgency. The study also uncovers communication dilemmas, including the challenge of balancing sustainability issues with leisure experiences and avoiding moral authority.DiscussionThe research emphasises the strength of combining communication models to appropriately reflect the multifaceted nature of wicked sustainability problems. By advancing the understanding of science communication on sustainability problems in a museum context, this research contributes to the development of innovative communication strategies for sustainability and ecosystem protection.
BackgroundThe patterns of group communication determine the way of learning, being involved, and knowledge building among students. The SWOT analysis, as an explanatory framework of the processes of learning, being involved, and building knowledge among students, is under-theorised in educational communication research in relation to the patterns of group communication.PurposeThis paper, based on a theoretical framework that includes SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis, was used to analyse group dynamics in the educational setting through archetypal patterns of group communication: Maypole, Round Robin, Hot Seat, and Free-Floating.ProcedureThe framework was synthesised from literature across various disciplines and was informed by focus group discussions with 28 male undergraduate students from three selected institutions in Delta State, Nigeria. The qualitative data were presented to empirically illustrate and modify the proposed conceptual framework, rather than testing it, and were analysed using thematic analysis.FindingsCommunication patterns act as structuring forces between individual agencies (Social Exchange Theory) and social structure (Social Identity Theory). The Maypole pattern concentrates on power and risks groupthink, whereas the Round Robin and Free-Floating patterns encourage equal participation and learning. Certain patterns strengthen and improve opportunities and diminish weaknesses and threats.ContributionThe article has a role in communication theory as it discusses the agency-structure dilemma in learning institutions and offers the teacher a diagnostic tool to strategically choose and apply patterns of communication in order to maximise learning. Empirical testing is provided with formal propositions.
This study examines how media organizations in the Gulf region mobilize knowledge economy principles and sustainable innovation frameworks to enhance organizational sustainability amid rapid digital transformation and the accelerating adoption of artificial intelligence (AI) technologies. Grounded in an integrated theoretical framework that synthesizes knowledge economy theory with sustainable innovation perspectives, the research investigates institutional readiness to transition toward knowledge-based models, AI's role in supporting operational and economic sustainability, structural knowledge transformations associated with digitization, and the ethical–social dimensions of technological change. A descriptive-analytical design was employed using a structured questionnaire administered to a purposive sample of 400 media professionals in Bahrain and the United Arab Emirates, yielding 390 valid responses (97.5% response rate). The instrument demonstrated acceptable internal consistency (Cronbach's α = 0.84) and a confirmed five-factor structure. Descriptive and inferential analyses—including chi-square tests, independent-samples t-tests, one-way ANOVA, and Spearman's rank-order correlations—were conducted to examine associations among the study's empirical axes. Findings reveal strong consensus regarding insufficient supportive regulatory policies for AI (69.0%) and significant concern about AI's potential impact on traditional jobs (62.8%). While respondents report high agreement that digital transformation improves knowledge management (82.9%) and that AI enhances efficiency and cost reduction (86.7%), persistent training gaps (75.7%), infrastructural constraints (61.3%), and substantial ethical uncertainty (only 39.0% agree that ethical considerations are adequately addressed) remain salient. Inferential analyses confirm a statistically significant negative association between perceived ethical adequacy and job-security anxiety, and a positive association between training availability and perceived organizational readiness. The paper proposes a three-pathway conceptual model—capability, governance, and cultural—that operationalizes the joint predictions of knowledge economy and sustainable innovation theory, and formulates practice-oriented recommendations for policymakers and media leaders.
Intangible Cultural Heritage (ICH) promotion on social media faces unique challenges, especially on feed-based, short- video platforms like Douyin. This study examines Intangible Cultural Heritage (ICH) videos on Douyin by decomposing the concept of a “hook strategy” into 24 distinct features and ranking their relative importance to user engagement. Our contribution is primarily a measurement framework that decomposes the otherwise loosely defined notion of a “hook” into concrete, reproducible textual and verbal features, together with a comparison of how strongly these features are associated with engagement in a large ICH corpus. We do not aim to predict engagement—which in this cultural-communication setting is inherently high-variance—but to compare the relative importance of these features. The findings reveal that importance concentrates in title-level and title–text features—Title Sentiment Score (TSS), Title–Text Semantic Similarity (TtSS), and a mutually-collinear semantic-magnitude block (Title and Text Semantic Density and their difference, TSD, tSD and TtSD). Because the random forest cannot separate this collinear block and does not generalize out of sample (held-out R2 ≈ 0), we test these associations in a covariate-adjusted regression layer that controls for length and corrects for multiple testing. Only Title Semantic Density survives: concise, focused titles (higher title-embedding magnitude) are associated with higher likes and comments after adjusting for length (β ≈ 0.30–0.35, FDR-p < 0.02). The remaining high-importance features—including title sentiment—are not robust to adjustment, and unadjusted correlations are weak because several measures co-vary with length. We read this gap between importance rank and adjusted association as a substantive result: most text-based hook signals do not survive proper adjustment, and engagement is dominated by exposure. The contribution is thus a reproducible measurement framework, one robust association (title focus), a construct-validity critique of norm-based semantic measures, and a cautionary negative result for text-only accounts of short-video engagement, pending experimental validation.
Heat-related illnesses (HRIs) are a growing public health concern, particularly in the context of global climate change. Vulnerable populations, such as agricultural and outdoor workers, face disproportionate risks of HRIs. This study investigates associations between social determinants of health (SDOH) profiles and prevention behaviors, trust in information sources, and frequency of information-seeking related to extreme heat among Florida residents. Using a Bayesian Profile Regression (BPR) approach, we identified four clusters based on six sociodemographic and contextual variables from an online survey administered to 500 Florida residents, including income, education, race/ethnicity, occupation, political orientation, and urban/rural status. Logistic regression models were then used to examine associations between these clusters and key HRI-related outcomes. Cluster 3, characterized by low income, low education, rural residency, agricultural work, and high proportions of Hispanic and Black people, showed elevated odds of low prevention and distrust in institutional information sources. Clusters 2 and 4 demonstrated more selective distrust of government-affiliated platforms. These findings underscore the need for targeted, culturally responsive health communication strategies tailored to the diverse SDOH profiles influencing heat-related vulnerability.Take-home messageSocial determinants of health profiles are associated with differences in prevention behaviors and trust in information sources during extreme heat. Populations characterized by lower income, lower education, and rural/agricultural backgrounds may experience greater heat-related vulnerability, highlighting the need for culturally responsive and targeted communication strategies.
