Spektrum Iran

Abstract Nearly five decades after Hubert Dreyfus underscored the importance of accounting for the social character of intelligence in the development of artificial intelligence, and notwithstanding artificial intelligence’s consolidation as an actant within the news media, practical implementations have progressed more rapidly than corresponding theoretical inquiry. Because the capacity for meaning-making within a social institution presupposes the socialization of a cognitive system, the socialization of artificial intelligence may be examined along a trajectory comparable to that of human natural intelligence. On this basis, the present article investigates the processes through which AI becomes socialized so as to assume a meaning-making role within a social institution such as the news media, addressing the central question: What constitutes socialized artificial intelligence? To this end, the study integrates the Dual-spacization of Intelligence with representation theory within a socio-organizational framework and adopts a retroductive theoretical approach to answering the research question. Within this analysis, social order is understood as an function of AI’s socialization process. The dual-spacization of the world consequently gives rise to a dual-spatial social order. The study’s findings suggest that AI may either be engineered to replicate existing forms of knowledge and entrenched social stereotypes in a manner analogous to human cognition, or be subject to social regulation that fosters an algorithmic rationality oriented toward the common good and the realization of a sustainable and just social order. Such an order depends on the opening of representational practices through reflexive engagement with social stereotypes, enabling transformations in representation and supporting increased diversity of identities. The contribution of this article lies in proposing an integrated model for understanding the mechanisms of AI socialization across meaning-producing social institutions. Furthermore, the model offers a comprehensive perspective on the socialization of both natural and artificial cognitive systems within the evolving structures of dual-spatial institutional social orders.

Abstract During times of crisis, the volume and speed of media data increase, making traditional analysis methods inadequate for information management and public awareness. Artificial intelligence, with its automation capabilities, sentiment analysis, crisis identification, and communication strategy provision, assists organizations in managing crises more effectively. This article, presented in a documentary style, seeks to explore the function of artificial intelligence in media data analysis and management during a crisis and how it enhances crisis management effectiveness in the information age. The theory utilized in this context is machine learning theory. By examining the challenges in media data analysis and crisis management, the article discusses the role of artificial intelligence in automating media monitoring, sentiment analysis, pattern and anomaly identification, and news trend prediction. It also delves into how AI can enhance crisis communication, detect misinformation, and offer more effective responses. Finally, the article addresses the challenges and limitations of using AI in media data analysis and crisis management, such as data bias, algorithmic transparency issues, and ethical considerations.During times of crisis, the volume and speed of media data increase, making traditional analysis methods inadequate for information management and public awareness. Artificial intelligence, with its automation capabilities, sentiment analysis, crisis identification, and communication strategy provision, assists organizations in managing crises more effectively. This article, presented in a documentary style, seeks to explore the function of artificial intelligence in media data analysis and management during a crisis and how it enhances crisis management effectiveness in the information age. The theory utilized in this context is machine learning theory.

Abstract The rapid proliferation of large language models and text-generative systems has precipitated not only technological transformation but also an epistemic reconfiguration of how meaning is produced and circulated. This paper diagnoses a specific risk facing Persian: the attenuation and potential displacement of its cultural-semantic horizon within globalized, predominantly English-language AI infrastructures. The objective is analytic and diagnostic: to delineate the conceptual grounds of 'semantic sovereignty' and to map the structural pathways through which contemporary AI practices endanger Persian meanings, metaphors, and hermeneutic traditions.

The study combines conceptual-philosophical analysis (philosophy of language, hermeneutics, phenomenology) with a critical reading of current AI training regimes and data ecologies. It employs analytic conceptual synthesis rather than empirical intervention: the analysis traces theoretical presuppositions (Wittgensteinian ‘meaning as use’, Gadamerian horizons, Davidsonian triangulation, Floridi’s information ethics) and maps them onto the material practices of dataset curation, model training, and platform mediation.

The paper identifies multiple, mutually reinforcing mechanisms by which AI systems produce semantic asymmetry: corpusic bias and representational scarcity; algorithmic translation that restructures non-English semantic networks into English-dominant vector spaces; infrastructural mediation that repositions Persian cultural artifacts as data-points divorced from their hermeneutic contexts; and epistemic filtering enacted by recommender and retrieval systems that privilege certain forms of explicability over opacity and singularity. Collectively these mechanisms instantiate what I term the ‘phenomenological extinction’ of a language’s world-disclosing power.

The phenomenon at stake is not mere lexical loss but an ontological impoverishment: a contraction of Persian’s capacity to disclose distinctive modes of being. Recognizing this risk requires conceptual clarity about semantic sovereignty as a diagnostic category. This paper stops short of prescribing remedial policies; instead, it aims to provide a rigorous philosophical staging of the problem so that subsequent scholarship and public discourse can assess the depth, modalities, and stakes of Persian’s semantic endangerment.

Abstract This study examines the multifaceted impact of artificial intelligence (AI) on interpersonal relationships within Iranian society, highlighting the cultural, social, and psychological challenges emerging from the rapid adoption of AI technologies. As tools such as virtual assistants, social media algorithms, and AI-driven communication platforms become embedded in daily life, they are reshaping patterns of interaction, emotional engagement, and cultural norms. Drawing on twelve semi-structured interviews analysed through a mixed-methods, qualitative-dominant approach, including thematic analysis, intercoder reliability checks, and cross-case comparison, the research identifies a dual narrative: AI enhances communication, productivity, and daily convenience, yet simultaneously undermines face-to-face engagement, emotional bonds, and traditional social practices central to Iranian culture.

Findings reveal growing concerns about weakened family and community ties, reduced social skills, dependency on intelligent systems, and generational gaps in digital adaptation. Participants also noted broader cultural shifts, including the rise of virtual lifestyles, threats to cultural identity, and increased social inequality driven by uneven access to AI tools. The study further identifies psychological risks such as loneliness, superficial online connection, diminished empathy, and the decline of emotional intelligence as individuals increasingly interact with algorithmic systems. At the societal level, privacy, data governance, and ethical challenges create additional pressures that shape public trust and relational dynamics. The study contributes to national and international debates on human–AI interaction by demonstrating how global technologies interact with local cultural contexts. It argues that balancing technological innovation with the preservation of Iranian social values is essential to ensuring that AI strengthens rather than erodes the foundations of meaningful human relationships.

Abstract This study investigates how Persian-speaking users on the social media platform X engage with generative artificial intelligence as a sociotechnical and discursive phenomenon. Drawing on a dataset of 24,215 Persian-language posts, we employ a multi-label topic modeling framework and affective profiling to analyze public discourse surrounding AI tools, their perceived implications, and normative judgments about their use. Rather than treating sentiment as a static indicator of opinion, we interpret affective expression as a communicative act shaped by platform incentives and cultural context. Our findings show that AI is positioned not only as a technical artifact but as a boundary object entangled with debates over expertise, ethics, and institutional legitimacy. The discourse is anchored in practical concerns—especially labor, education, and tool comparisons—but frequently extends into culturally specific narratives about risk, fairness, and epistemic authority. Emotionally, the conversation is marked by pragmatic positivity, critical intensity, and a sizable neutral band reflecting orientation rather than evaluation. This study contributes to ongoing debates in communication, AI ethics, and platform studies by offering a non-Anglophone, culturally grounded analysis of how publics perform vernacular governance over emerging technologies. Emotionally, the conversation is marked by pragmatic positivity, critical intensity, and a sizable neutral band reflecting orientation rather than evaluation. This study contributes to ongoing debates in communication, AI ethics, and platform studies by offering a non-Anglophone, culturally grounded analysis of how publics perform vernacular governance over emerging technologies. Drawing on a dataset of 24,215 Persian-language posts, we employ a multi-label topic modeling framework and affective profiling to analyze public discourse surrounding AI tools, their perceived implications, and normative judgments about their use.

Keywords: artificial intelligence; sentiment analysis; digital publics; vernacular governance; Persian discourseKeywords: artificial intelligence; sentiment analysis; digital publics; vernacular governance; Persian discourse

Abstract This study examines the relationship between data bias, algorithmic ethics, and the social consequences of digital hermeneutics. With the expanding influence of artificial intelligence in interpretive domains particularly in the interpretation of religious and philosophical texts the issue of data neutrality and algorithmic objectivity has become a fundamental concern. The findings of this research, using an analytical–explanatory approach, demonstrate that training data, contrary to common assumptions, are not neutral; rather, they embody cultural values and presuppositions that are reproduced within algorithmic processes. Such reproduction can lead to semantic simplification, the elimination of interpretive diversity, and even distortion of the content of sacred texts.

Drawing upon a hermeneutical perspective, the present article highlights the necessity of distinguishing between “human pre-understanding” and “machine data,” showing that the absence of awareness, critical reflexivity, and lived experience in algorithms prevents the attainment of authentic understanding. Moreover, the study indicates that the social implications of this situation extend beyond textual interpretation, posing threats to privacy, exacerbating social inequalities, and weakening cultural diversity. Ultimately, the article emphasizes that digital hermeneutics can be constructive only when, alongside the technical capacities of artificial intelligence, ethical principles, religious oversight, and the preservation of interpretive traditions are upheld.

Introduction

The emergence of artificial intelligence and machine-learning algorithms has opened new horizons within the humanities and the interpretation of texts horizons that simultaneously entail fundamental opportunities and profound threats. Digital hermeneutics, as an interdisciplinary approach situated at the intersection of information ethics, the philosophy of understanding, and modern technologies, raises a critical question: how is the position of the human being, as the interpretive agent, to be redefined in the age of algorithms? Although large language models possess remarkable capacities for organizing data and generating text, the questions concerning the validity, objectivity, and ethical soundness of their outputs remain unresolved.

Abstract Public trust in media constitutes a core component of social capital and communicative legitimacy, yet it is increasingly challenged by the rapid integration of artificial intelligence and synthetic media into news production and distribution processes. This study explores alternative futures of public trust in media in the age of artificial intelligence and develops scenario-based insights for Iran toward the horizon of 2036. Adopting a futures-studies approach, the research employs a mixed-methods design that combines environmental scanning and a systematic review of academic and policy sources (2018–2025) with a two-round Delphi consultation involving fifteen experts in media, artificial intelligence, and governance. Structural analysis using the MICMAC method was applied to examine influence–dependence relationships among key variables, leading to the identification of media transparency and the quality of AI regulation as the two critical uncertainties shaping future trajectories of public trust. Based on these axes, four alternative scenarios were developed—Smart Trust, Total Distrust, Islands of Trust, and Imposed Trust—each illustrating a distinct configuration of governance choices, technological use, and audience responses. The findings demonstrate that future patterns of public trust are not technologically deterministic but are primarily driven by institutional transparency, regulatory arrangements, and governance decisions. The study concludes that strengthening accountable AI governance, enhancing media transparency, and investing in media literacy among audiences are essential for steering Iran’s media ecosystem toward a sustainable and trust-based future.

Abstract This study investigates how Persian-speaking users on X interpret and emotionally respond to DeepSeek, a Chinese-developed large language model. Drawing on a curated corpus of 1,112 posts collected from Iranian users, the research employs a mixed-method approach incorporating sentiment analysis, topic modeling, and co-occurrence network analysis. The findings reveal a layered discursive landscape in which DeepSeek serves not merely as a technological product but as a symbolic site for negotiating issues of geopolitical alignment, epistemic trust, and technological aspiration. Six major affective orientations—neutrality, skepticism, hope, pride, anxiety, and dismissiveness—structure user engagement with the model, reflecting ambivalent yet politically informed responses. Thematic analysis identified eight recurring topics, including performance comparisons, Chinese sovereignty, AI ethics, and cultural identity, which often co-occurred in complex rhetorical configurations. These results suggest that Iranian users deploy DeepSeek as a proxy to reflect on domestic technological constraints, platform politics, and the shifting contours of global AI hegemony.

Abstract Objective: Artificial intelligence (AI) is increasingly used as a key tool in redefining digital branding and customer engagement. It encompasses techniques and methods that businesses leverage to create brand value, enhance the effectiveness of customer interactions, and improve marketing strategies. Analysis of interviews indicates that AI applications—through data analysis, advanced algorithms, modeling, and other techniques—bring significant transformations in digital branding processes, while also presenting specific challenges and opportunities. Accordingly, this study focuses on identifying AI techniques, persuasive effects, transformations, and challenges associated with AI implementation in digital marketing and customer engagement.

Method: To address the research questions, this study employed a qualitative, field-based approach. Seventeen experts in AI and digital branding were purposefully selected and studied through semi-structured interviews. Participant selection focused on expertise, professional experience, and practical familiarity with AI applications in digital branding. The interviews aimed to explore experts’ experiences, perceptions, and insights regarding AI’s role and functions in branding processes. The interview data were analyzed using thematic analysis. Initially, codes were extracted from interview transcripts, which were then categorized and aggregated to identify sub-themes and, ultimately, the main research themes.

Results: The study found that the main themes regarding AI techniques in digital branding include scalable computational algorithms, machine learning, reinforcement learning algorithms, search algorithms, operational automation, recommender algorithms, data interpretation and processing, human-computer interaction algorithms, and AI-based platforms. Regarding AI-driven transformations in digital branding, themes such as dynamic digital marketing, transformation in distribution channels, personalized brand communications, dynamic pricing, adaptive business strategies, enhanced cybersecurity, development of internal and external databases, improved customer experience, strengthened brand positioning, development of organizational processes and systems, enhanced decision-making quality, brand globalization, formation of digital business models, and value creation were identified.

Abstract Large Language Models (LLMs) are usually considered neutral technological advancements. However, critical digital studies increasingly emphasize the need to challenge their potential to perpetuate colonial power structures in cyberspace. This paper argues that LLMs function as powerful apparatuses of digital neocolonialism. It aims to diagnose this phenomenon within the field of literary AI and to propose a decolonial framework for its future development. This study demonstrates how the protocols of extracting and processing data privilege Western epistemologies in a systematic manner. Then, it develops a conceptual framework for the praxis of decolonial AI based on the principles of reciprocity and epistemic justice. The analysis reveals that the extractivist data collection utilized by dominant LLMs treats cultural and linguistic data as territory for appropriation, privileging the Western literary canon and erasing marginalized languages and traditions. This has led to linguistic homogenization and epistemic injustice as well as the imposition of aesthetic standards of the global West. In response, the proposed decolonial framework has necessitated a paradigm shift from extraction to reciprocity, which involves community-led data governance. Furthermore, AI should be used as a collaborative, co-creative tool by literary writers and researchers. As a further decolonial step, Eurocentric evaluative criteria in this field must be reformed in concrete ways. The decolonial approach advanced in this paper, seeks to fundamentally reposition literary AI. The ultimate goal of this repositioning is to foster a pluriversal aesthetic and epistemic framework.

Abstract Humans anthropomorphize digital entities, such as Generative Artificial Intelligence (GAI), assigning them human-like physical traits, mental states, or social characteristics, including gender. GAI, as a sociotechnical actor, both reflects and shapes the society that produces it. Similarly, the intersections of GAI and gender are mutually co-constitutive. Gender is embedded, reproduced, enacted, materialized, and embodied in AI technologies. The current research explores anthropomorphism and the gendering of GAI from a social constructionist perspective, examining how individuals consciously and unconsciously adopt stereotypical gendered expectations when anthropomorphizing GAI. An embedded mixed-methods design was employed, with quantitative data nested within a predominantly basic qualitative research approach. Qualitative and quantitative data were collected simultaneously via purposive and convenience sampling, and sixty-seven Iranian participants completed the online questionnaire. The study began with an autoethnographic vignette. The quantitative strand followed the logic of Q methodology, identifying distinguishing items by treating participants as variables in the analysis. Qualitative data were analyzed using thematic analysis. Over half of the participants did not assign a gender or name to GAI, while roughly half of the remaining participants assigned a variable gender (male, female, or genderless), the remainder attributed a fixed gender, which was predominantly male. Many participants did not anthropomorphize GAI, emphasizing its machinic nature, whereas other participants’ responses revealed that human-like attachments, gender assignments, naming practices, and the ways these anthropomorphic exercises are shaped by GAI use mirror broader cultural norms, indicating that perceived gender in GAI is socially enacted rather than intrinsic. Since emotional bonds with increasingly humanized GAI chatbots can lead to negative or positive outcomes, GAI literacy is necessary. Policymakers and educational institutions should devise initiatives to raise GAI literacy, and that GAI corporations adopt self-regulatory measures to protect users.

Abstract This study explores how language influences sentiment classification in outputs generated by a multilingual large language model (LLM), Grok. Anchored in Langdon Winner’s theory of technological politics—which holds that technologies are inherently non-neutral and embed structural biases—this study examines whether sentiment distributions vary systematically across languages even under a fully harmonized analytic pipeline. The analysis draws on a corpus of 4,799 posts (English: n = 2,399; Persian: n = 2,400) generated using identical prompts. Sentiment outputs were mapped onto a common three-category schema (Negative, Neutral, Positive), and analyses incorporated both discrete class assignments and continuous probability scores. To account for superficial textual variation, structural characteristics—including sentence, word, and character counts—were computed and incorporated as controls. The results demonstrate a robust cross-linguistic divergence in sentiment patterns. English-language outputs are predominantly clustered around Neutral classifications and exhibit comparatively lower affective intensity, whereas Persian-language outputs display a pronounced shift toward Positive sentiment accompanied by greater dispersion. Crucially, these differences remain statistically significant after controlling for structural features, suggesting that language affiliation, rather than text length or segmentation, constitutes the principal correlate of observed sentiment variation. At the probability level, English distributions exhibit a tighter concentration around neutral probabilities, while Persian distributions are flatter and more positively skewed, with higher intensity indices. These results have significant implications for multilingual sentiment analysis and LLM auditing. If language effects are not explicitly modeled and calibrated, comparative analyses may conflate linguistic variation with affective intent, leading to distorted inferences about tone, stance, or emotional valence. The study underscores the importance of reporting both label and probability metrics, adopting language-specific calibration protocols, and treating language as a first-order measurement dimension in cross-lingual content analysis.