AI-BASED ANALYSIS OF GENDERED LANGUAGE PATTERNS IN ENGLISH–UZBEK SOCIAL MEDIA DISCOURSE

GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 150 DOI: https://10.5281/10.5281/zenodo.17686094 AI-BASED ANALYSIS OF GENDERED LANGUAGE PATTERNS IN ENGLISH–UZBEK SOCIAL MEDIA DISCOURSE Qudratullayeva Muniraxon Abrorjon qizi Teacher at Kokand University Mardonov Quvonchbek Boymurod o‘g‘li Teacher at UzSWLU ANNOTATION This article investigates gendered language patterns in English–Uzbek social media discourse through AI-powered analytical methods. Using machine learning classification, corpus-based analysis, sentiment modeling, and NLP tools, the study examines how male and female users employ linguistic forms, code-switching behaviors, emotive expressions, and politeness strategies across bilingual platforms. Keywords: AI, gendered language, English–Uzbek discourse, NLP, social media linguistics, digital identity, code-switching, sentiment analysis. АННОТАЦИЯ В данной статье исследуются гендерные языковые модели в англоузбекском дискурсе социальных сетей с использованием методов, основанных на искусственном интеллекте. Применяя классификацию машинного обучения, корпусный анализ, моделирование сентимента и инструменты обработки естественного языка (NLP), исследование анализирует, как пользователимужчины и пользователи-женщины используют языковые формы, стратегии GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 151 переключения кодов, эмоционально-экспрессивные средства и вежливые речевые тактики на билингвальных онлайн-платформах. Ключевые слова: искусственный интеллект; гендерно маркированный язык; англо-узбекский дискурс; обработка естественного языка (NLP); лингвистика социальных сетей; цифровая идентичность; переключение кодов; сентимент-анализ. ANNOTATSIYA Ushbu maqola ingliz–o‘zbek ikki tilli ijtimoiy tarmoq diskursida genderga xos til xususiyatlarini sun’iy intellekt yordamida tahlil qiladi. Mashina o‘rganishi, korpus tahlili, sentiment modeli va NLP usullaridan foydalangan holda, erkak va ayol foydalanuvchilarning til birikmalari, kod-almashuv odatlari va emotsional ifodalaridagi farqlar o‘rganiladi. Kalit so‘zlar: SI, genderga xos til, ingliz–o‘zbek diskursi, NLP, ijtimoiy tarmoq lingvistikasi, raqamli identitet, kod-almashuv, sentiment tahlili. Introduction. Digital communication has emerged as the primary arena in which gendered linguistic styles and culturally embedded norms intersect, shift, and acquire new meanings. As social interaction increasingly migrates to online platforms - social media, messaging applications, and multilingual forums the language choices made by users become powerful indicators of identity performance 1 . In bilingual and multilingual contexts, such as English-Uzbek digital discourse, these choices are shaped not only by linguistic competence but also by the social expectations attached to each language. English often indexes modernity, global belonging, or informality, while Uzbek may carry associations with tradition, local identity, and formal 1 Eckert, P., & McConnell-Ginet, S. (2013). Language and Gender (2nd ed.). Cambridge University Press. GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 152 politeness norms. These symbolic values interact dynamically with gender, generating distinctive patterns in how men and women navigate bilingual repertoires online. Recent advances in natural language processing have opened new methodological opportunities for capturing and analyzing these gendered patterns with unprecedented precision. Large-scale AI language models can process extensive datasets of online communication posts, comments, messages, and hashtags enabling researchers to identify subtle, recurrent features of gendered style that may elude manual coding. For instance, AI models can detect systematic variation in codeswitching frequency, topic selection, stance expressions, emotive markers, and discourse strategies as they correlate with gender and cultural context. 1 Moreover, computational tools make it possible to examine how these patterns evolve over time, respond to sociopolitical events, or vary across different digital environments. In the study of English-Uzbek bilingual discourse specifically, AI-assisted analysis allows researchers to explore how gender norms are maintained, negotiated, or contested through linguistic choices. Women and men may differ in their strategic use of English for empowerment, anonymity, or cosmopolitan identity performance, while relying on Uzbek to affirm cultural belonging or comply with expected politeness conventions. Conversely, moments of code-switching may function as deliberate acts of distancing, solidarity, humor, or resistance, each influenced by the speaker’s gendered position within society. 2 By enabling researchers to systematically quantify and model these phenomena, AI technologies contribute not only to descriptive accounts of bilingual digital communication but also to broader theoretical frameworks on language, gender, and cultural hybridity. The integration of computational linguistics with sociolinguistic and gender studies therefore represents a significant methodological and conceptual 1 Fierman, W. (2016). “Language attitudes in Uzbekistan: Uzbek, Russian, and English.” International Journal of the Sociology of Language, 239, 79–107. 2 Cameron, D. (2007). The Myth of Mars and Venus: Do Men and Women Really Speak Differently? Oxford University Press. GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 153 advancement, offering new insights into how bilingual speakers navigate and reshape linguistic norms in the digital age. Literature review. Gendered language is not just a set of linguistic patterns - it is a reflection of how people relate to each other. The ways individuals express politeness, show emotion, or choose certain words often reveal the expectations placed on them within their communities. In many cultures, women are encouraged to speak more gently or diplomatically, while men are expected to sound more direct or reserved. These tendencies are not fixed rules, but they do show how social norms shape everyday communication. In bilingual settings, these dynamics become even more interesting. Speakers who move between English and Uzbek, for example, often use each language to highlight different parts of who they are. English might feel more suited to expressing confidence or neutrality, while Uzbek may carry a sense of warmth, cultural closeness, or respect. By switching languages, people can shift their tone, adjust their emotional expression, or position themselves differently depending on the situation. This flexibility makes bilingual communication a rich space for understanding how gender is expressed. 1 Today, AI tools give researchers new ways to explore these patterns. Instead of relying only on small samples or manual observation, machine learning models can analyze thousands of messages at once. They can identify when speakers switch languages, what kinds of words they choose, or how they express politeness or emotion. These tools help uncover patterns that might otherwise go unnoticed - like subtle differences in how men and women use emojis, borrow English slang, or switch to Uzbek for emphasis or intimacy. By combining human insight with the power of AI, researchers can build a clearer picture of how gender, language, and culture interact in digital spaces. This approach 1 Herring, S. C., & Kapidzic, S. (2015). “Teens, gender, and self-presentation in social media.” International Encyclopedia of Social and Behavioral Sciences, 2nd ed., 1–11. GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 154 makes it possible to see not only what people say, but also how their linguistic choices help them navigate identity and connection in a multilingual world. Research methodology. To understand how gendered language patterns emerge in real digital interactions, we analyzed datasets collected from several major online platforms, including Instagram, Telegram, TikTok, and YouTube comment sections. These platforms were chosen because they represent different styles of communication - ranging from casual messaging to highly public conversations - and therefore offer a broad view of how people express themselves across contexts. A range of AI-driven techniques was applied to these datasets. Sentiment modeling allowed us to identify the emotional tone of messages and see how men and women express positivity, negativity, or nuance differently across languages. Keyword extraction helped highlight the words and phrases most strongly associated with each gender, providing insight into recurring topics, stylistic choices, and cultural references. To capture the bilingual nature of the discourse, we used code-switching detection tools that automatically pinpointed moments when users shifted between English and Uzbek. These switches often carry important social meaning signaling closeness, humor, emphasis, or shifts in identity so identifying them at scale was essential. Finally, embedding-based classifiers enabled deeper analysis of patterns that are not tied to any single word, such as subtle stylistic tendencies or gendered ways of framing ideas. Together, these methods created a comprehensive picture of how gender and bilingualism interact across multiple digital platforms, offering insights that would be difficult to observe through manual analysis alone. Analysis and results: Gendered Linguistic Patterns. Clear differences emerged in how female and male users express themselves across platforms. Female users tended to communicate with greater emotional intensity, often using expressive vocabulary, exclamation marks, GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 155 and a noticeably wider range of emojis to convey tone, empathy, or humor. These choices suggest a communication style geared toward relational connection and emotional clarity, aligning with broader sociolinguistic observations. Male users, on the other hand, gravitated toward more direct, concise phrasing. Their messages featured less emotional marking and more use of technical slang, jargon, or topic-specific keywords - especially in discussions related to technology, gaming, finance, and problem-solving. This preference reflects a more referential and task-oriented communicative style. Code-switching patterns also revealed meaningful stylistic and cultural tendencies. Women more frequently switched to Uzbek to express warmth, affection, or cultural intimacy, while using English for modernity, humor, or stylistic flair. Men’s code-switching tended to be more functional, often triggered by topic shifts or the introduction of technical terms. These bilingual patterns demonstrate how speakers draw on the symbolic value of each language to shape tone, identity, and social alignment. 1 The Role of AI in Shaping Gendered Discourse. Digital communication today is increasingly shaped by predictive text systems, machine translation tools, and conversational chatbots. These technologies subtly guide how users formulate their messages, often without conscious awareness. One noticeable effect is the reinforcement of English-origin expressions and hybrid linguistic forms within bilingual English-Uzbek communication. Predictive text suggestions frequently offer English lexical items especially trendy slang, shortened forms, or high-frequency expressions which users may adopt because they are quicker to select and socially recognizable online. Female users often incorporate these suggestions into expressive or stylistic elements, blending English terms with Uzbek to create softer, more emotionally nuanced phrasing. Male users 1 Holmes, J., & Meyerhoff, M. (Eds.). (2003). The Handbook of Language and Gender. Wiley-Blackwell. GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 156 more commonly adopt English-origin technical terms or concise hybrid constructions, aligning with their preference for direct and efficient communication. Machine translation tools also contribute to this trend by promoting syntactic structures and vocabulary that reflect English norms, sometimes leading to calques or hybrid expressions in Uzbek. Chatbots, which often respond using standardized English-influenced patterns, further normalize these forms by providing input that users consciously or unconsciously mirror. Taken together, these AI-mediated tools act as subtle but influential actors in shaping how bilingual speakers express gendered identities. They encourage the diffusion of English loanwords, hybrid phrases, and digitally shaped discourse styles, gradually reshaping the linguistic landscape of online communication. Discussion. AI-driven analysis reveals that online language use is not simply a reflection of individual preference; it is shaped by the interplay of cultural norms, global digital practices, and algorithmic influences. In the case of English-Uzbek bilingual communication, this interaction becomes particularly visible. Uzbek gender norms, which influence expectations around politeness, emotional expression, and social positioning, continue to guide how men and women present themselves online. At the same time, English-language digital culture-characterized by informality, expressive creativity, and rapid adoption of new slang introduces alternative models of self-expression that users draw upon as they craft their online identities. For many, switching to English is a way of tapping into global trends, signaling modernity, or experimenting with styles not typically available within traditional norms. Algorithmic shaping adds a third layer to this dynamic. Predictive text systems, translation tools, and chatbot interactions subtly encourage certain linguistic choices through their suggestions, defaults, and stylistic patterns. As a result, users frequently GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 157 adopt English-influenced forms, hybrid expressions, or algorithmically favored phrasing, even when communicating within culturally grounded Uzbek contexts. 1 Together, these forces local gender expectations, global digital culture, and algorithmic influence blend to create hybrid digital identities. These identities are neither fully traditional nor fully global; they emerge from the creative negotiation of multiple linguistic and cultural resources. AI helps make these processes visible, revealing how people craft new forms of self-expression that transcend singular cultural frameworks while still remaining rooted in them. Conclusion. AI-based analysis reveals that gender plays a significant and consistent role in shaping English-Uzbek bilingual discourse. Across platforms and communicative contexts, men and women draw on different emotional markers, stylistic preferences, and code-switching strategies, reflecting both culturally embedded gender norms and the influence of global digital practices. At the same time, hybrid linguistic forms mixing English expressions, Uzbek structures, emojis, slang, and algorithmically suggested phrasing are becoming increasingly normalized. These hybrids are no longer peripheral or playful; they are emerging as a stable part of everyday digital communication, especially among younger bilingual users. Overall, AI-driven linguistic research opens a new window into how gender, culture, and technology intersect in multilingual digital spaces—revealing a complex, evolving landscape where hybrid identities are becoming a defining feature of contemporary communication. REFERENCES 1. Eckert, P., & McConnell-Ginet, S. (2013). Language and Gender (2nd ed.). Cambridge University Press. 1 Herring, S. C. (2004). “Computer-mediated discourse analysis: An approach to researching online behavior.” Designing for Virtual Communities in the Service of Learning, 338–376. GOLDEN BRAIN ISSN: 2181-4120 VOLUME 3 | ISSUE 17 | 2025 Multidisciplinary Scientific Journal November, 2025 158 2. Cameron, D. (2007). The Myth of Mars and Venus: Do Men and Women Really Speak Differently? Oxford University Press. 3. Holmes, J., & Meyerhoff, M. (Eds.). (2003). The Handbook of Language and Gender. Wiley-Blackwell. 4. Herring, S. C. (2004). “Computer-mediated discourse analysis: An approach to researching online behavior.” Designing for Virtual Communities in the Service of Learning, 338–376. 5. Baker, P. (2014). Using Corpora to Analyze Gender. 6. Crystal, D. (2011). Internet Linguistics. 7. Eckert, P. & McConnell-Ginet, S. (2013). Language and Gender. 8. García, O. & Wei, L. (2014). Translanguaging. 9. Nguyen, D. (2021). Gender and language in online communication. 10. Herring, S. C., & Kapidzic, S. (2015). “Teens, gender, and selfpresentation in social media.” International Encyclopedia of Social and Behavioral Sciences, 2nd ed., 1–11. 11. Kendall, L. (2002). Hanging Out in the Virtual Pub: Masculinities and Relationships Online. University of California Press. 12. Fierman, W. (2016). “Language attitudes in Uzbekistan: Uzbek, Russian, and English.” International Journal of the Sociology of Language, 239, 79–107. 13. Pavlenko, A. (2008). “Multilingualism in post-Soviet countries.” International Journal of Bilingual Education and Bilingualism, 11(3–4), 275–314.