WHEN AI ISN’T ENOUGH: THE UI/UX EXCEPTION IN AI-SUPPORTED LEARNING AT JAPAN DIGITAL UNIVERSITY

Аннотация

This paper explores developing culturally-aware AI models for Uzbek language education, emphasizing the integration of national values and oral traditions. Traditional language learning often neglects crucial cultural nuances, leading to proficient but culturally inept speakers. We argue that AI in language education must embed cultural understanding into its core, moving beyond homogenous datasets that risk bias. The text highlights Uzbek's rich linguistic and cultural landscape, rooted in values like hospitality and oral heritage (proverbs, epics). It outlines approaches for integrating these elements into AI through cultural annotation, knowledge graphs, and interactive modules. Challenges like data scarcity and algorithmic bias are discussed, alongside ethical considerations and evaluation frameworks to ensure authentic, effective, and culturally sensitive AI-powered language learning. Key words: AI language learning, Uzbek culture, national values, oral traditions, cultural sensitivity I. Introduction. In today’s increasingly interconnected world, true communication requires more than mastering grammar and vocabulary—it demands an in-depth appreciation of cultural norms and sensitivities.[1, 2] Traditional language instruction, and even many AI-driven systems, often overlook this essential cultural layer, rendering learners proficient in structure but deficient in real-world communication skills.[1] Many such methods reduce language learning to technical memorization, detached from the cultural environment where language naturally operates. As a result, students often struggle to apply their knowledge in meaningful, culturally appropriate ways.[1] To move beyond these limitations, artificial intelligence used in language education must treat culture not as an optional supplement but as an essential component. Only by embedding cultural understanding into the very architecture of AI systems can we foster learners who are not just fluent but also capable of cross-cultural engagement. Recent developments in AI and Natural Language Processing (NLP) have already revolutionized education through adaptive platforms, chatbots, and AI tutors that deliver instant feedback and personalized instruction.[2, 3] These 70 ACADEMIC RESEARCH IN MODERN SCIENCE International scientific-online conference tools relieve educators of repetitive tasks and help maintain student engagement through interactive, gamified content.[2] State-of-the-art models like GPT-3 and BERT have become central to such innovations, excelling in language generation, contextual understanding, and translation.[1] However, when these models are built on culturally homogenous datasets—especially those heavily centered on Western norms—they often fail to interpret context- sensitive expressions or reflect diverse cultural values.[1, 4] This issue, sometimes termed “stereotype leakage,” arises when biased cultural perspectives in training data surface in unrelated linguistic contexts, leading to miscommunication or the unintentional reinforcement of stereotypes.[4] Spoken by millions throughout Central Asia, Uzbek is a cornerstone of national identity in Uzbekistan and a vessel for centuries of cultural legacy.[5, 6] It belongs to the Karluk branch of Turkic languages and features rich dialectal diversity—Northern variants bear traces of Russian, while Southern dialects incorporate Persian influences.[6, 7] These dialectical differences are more than linguistic—they reveal regional identities and historical trajectories, which in turn shape perceptions and expressions.[7] For instance, loanwords from Persian and Russian not only enrich the language but also reflect past political and cultural alliances.[7] Uzbek culture itself is rooted in time-honored values like hospitality, reverence for elders, and a strong emphasis on family and collectivism.[5, 8, 9, 10] These principles permeate language use, from the ubiquity of kinship terms (e.g., aka, opa, uka, singil) in social address to the customary use of indirect praise and formal greetings.[11] Folk expressions, parables, and epics serve as vehicles for transmitting this cultural ethos, making them essential teaching tools.[8, 10] These oral traditions don’t just preserve stories—they encode ethical beliefs, shared history, and linguistic subtleties that define Uzbek identity.[10] II. Main Part. What Is Culturally-Aware AI and Why Does It Matter in Language Learning? Culturally-aware AI refers to systems capable of discerning and adapting to the cultural contexts and values of the people interacting with them.[1] Unlike traditional AI models focused on syntactic accuracy, these systems aim for culturally meaningful communication by recognizing idioms, social cues, and context-specific behavior.[1, 4] For example, where Western design philosophy might treat AI as a tool to optimize productivity, Eastern perspectives may embrace AI as a social companion that complements human relationships.[1] 71 ACADEMI