Artificial intelligence tools as a means of enhancing EFL students‘ idiomatic competence

Maqolada muallif universitet bosqichida ingliz tilini chet tili sifatida o‘rganayotgan talabalarga idiomalarni o‘qitishda sun‘iy intellekt (SI) vositalarini auditoriya mashg‘ulotlariga joriy etish masalasini muhokama qiladi. Idiomalar ko‘chma ma‘nolari, madaniy nozikliklari va qo‘llanish kontekstlari tufayli talabalar uchun so‘z boyligining ayniqsa murakkab qismi hisoblanadi. Maqolada idiomalarni o‘qitishda turli SI vositalaridan, jumladan, yirik til modellariga asoslangan generativ chatbotlar, SI asosidagi matn yozish vositalari, matndan tasvir yaratuvchi modellar, nutq texnologiyalari va SI yordamida ishlaydigan mashq generatorlaridan foydalanish o‘rganiladi. Nazariy mulohazalar va o‘quv jarayonini loyihalash asosida muallif SI yordamida idiomalarni o‘qitishning to‘rt bosqichli yondashuvini taqdim etadi, topshiriqlar tasnifini taklif qiladi va SI yaratgan kontentni tanqidiy baholash mezonlarini belgilaydi. Talabalarning ona tiliga alohida e‘tibor qaratiladi: ingliz tilidagi oziq-ovqat bilan bog‘liq idiomalar SI vositalari yordamida ularning qoraqalpoq tilidagi muqobillari bilan qiyoslanadi, taklif etilgan muvofiqliklarning to‘g‘riligini esa o‘qituvchi tekshiradi. SI vositalari tanqidiy fikrlash va talabalar mustaqilligini rivojlantirishni rag‘batlantiradigan, o‘qituvchi tomonidan ishlab chiqilgan o‘quv dasturining tarkibiy qismi bo‘lganda idiomalarni o‘zlashtirishga ko‘maklashishi mumkin, degan fikr ilgari suriladi.

R E S E A RC H F O C US xalqaro ilmiy jurnali hamkorligida

QR-kod maqolaning DOI sahifasini ochadi SAHIFALAR TIL 762–767 Ingliz IQTIBOS UCHUN Madina Amangalievna Kalabaeva. Artificial intelligence tools as a means of enhancing EFL students’ idiomatic competence // «O‘zbekiston olima ayollarining mamlakat taraqqiyotiga qo‘shayotgan hissasi» mavzusidagi xalqaro ilmiy-amaliy konferensiya materiallari (Nukus, 2026-yil 9–10- oktyabr). – Nukus, 2026. – B. 762– 767. – DOI: https://doi.org/10.66073/ ReFocus-Conf/0910151

«O‘zbekiston olima ayollarining mamlakat taraqqiyotiga qo‘shayotgan hissasi» XALQARO ILMIY-AMALIY KONFERENSIYA NUKUS, 9–10-OKT YABR 2026

ARTIFICIAL INTELLIGENCE TOOLS AS A MEANS OF ENHANCING EFL STUDENTS’ IDIOMATIC COMPETENCE Madina Amangalievna Kalabaeva PhD in Pedagogical Sciences, Associate Professor Karakalpak State University named after Berdakh Associate Professor, Department of English Linguistics Email: mkalabaeva70@gmail.com

Abstract: In this article, the author discusses the integration of artificial intelligence (AI) tools in the classroom for teaching idioms to English as a foreign language students at a university level. Idioms are a particularly challenging part of vocabulary for students, given their figurative meanings, the cultural nuances, and the contexts in which they are used. The article explores the uses of various AI tools in idiom pedagogy, including generative chatbots based on large-scale language models, AI-based writing tools, text-to-image models, speech technologies, and AI-driven exercise generators. Through theoretical reflection and instructional design, the author introduces a four-step approach to teaching idioms with AI, offers a classification of tasks, and sets forth criteria for the critical assessment of AI-generated content. The exploration of students’ first language is an area of specific concern: English food-related idioms are compared with their Karakalpak counterparts using AI tools, and the teacher checks the suggested equivalences for accuracy. It is suggested that AI tools can support idiom acquisition when they are part of an instructor-designed curriculum that encourages critical thinking and the development of student autonomy. Keywords: artificial intelligence; idiomatic competence; idioms; EFL; large language models; chatbots; contrastive analysis; Karakalpak learners; learner autonomy; digital education. Introduction The fast-evolving field of artificial intelligence (AI) is now recognized as a crucial aspect of what it means to be “educated”. In light of widespread dissemination of various generative AI technologies in late 2022, students and teachers worldwide now have access to tools that can explain terminology, create content, facilitate conversations, and even provide immediate feedback. This trend has received attention in policy in Uzbekistan. Specifically, in 2023, the government of Uzbekistan outlined the “Uzbekistan – 2030” Strategy, which included digitalizing the education system and increasing access to education. It is expected that the trend will continue, with the government of Uzbekistan, for example, recently adopting the Strategy for the Development of Artificial Intelligence Technologies until 2030, which seeks to expand the application of AI across all sectors, including education. Against this backdrop, this research paper aims to explore the issue of how AI can be used in a robust and responsible manner in the teaching and learning of foreign languages.

This discussion will focus on the teaching of idioms as a potential application of AI in language instruction. An idiom is a fixed metaphorical expression (e.g., to spill the beans, a piece of cake) whose meaning cannot be deduced from the individual words that comprise it. Idioms are a type of formulaic language, which is known to be central to fluent, authentic communication (Wray, 2002). According to NUKUS · 2026

Liu (2008), idiomatic expressions are common in both spoken and written English; as such, acquisition of idioms is considered a component of advanced language proficiency. Accordingly, we may view idiomatic competence as the ability to recognize, comprehend, and use idiomatic expressions in a manner that is sensitive to context, register, and culture. X A L Q A R O I L M I Y- A M A L I Y K O N F E R E N S I YA

As with other types of formulaic language, these phrases can pose significant difficulties for learners. Learners revert to L1-based idiom interpretation and production, with idioms closest to L1 forms being the most accessible and idioms without L1 counterparts being the hardest and the ones most likely to be avoided.[5. 287-304]. Learners also attempt to derive meanings for unknown idioms from contextual clues or by interpreting them literally.[2] Cognitive linguists have found that a large number of idioms are based on conceptual metaphors [8.9] and that teaching the conceptual base for the idiom and the real-world imagery behind the figurative meaning can aid in comprehension and that memory can be improved through the use of both verbal and visual cues.[1.] Language education AI research is « O ‘ Z B E K I S T O N O L I M A AY O L L A R I N I N G M A M L A K AT T A R A Q Q I Y O T I G A Q O ‘ S H AY O T G A N H I S S A S I »

accelerating. Huang, Hew, and Fryer, reviewing chatbot-supported language learning, found advantages such as immediate support, decreased anxiety, and enhanced practice possibilities, as well as drawbacks related to the novelty effect and cognitive overload. [4.237-257] ChatGPT can clarify word/phrase meanings in context, produce examples and dialogues, and tailor explanations to student’s proficiency, but mention issues of false or inaccurate content and importance of teacher/student’s digital literacy.[7.537-557] Godwin-Jones explores AI writing assistant’s integration as a writing partner, and Kasneci et al. highlight the potential benefits of generative language models for education and the threats posed by bias, hallucinations, and plagiarism. UNESCO (2023) advocates a human-centred approach to AI in which AI aids but does not substitute for teacher. Conversely, using AI tools for developing idiomatic competence has been relatively underexplored, especially with learners whose L1 is digitally underrepresented, for instance Karakalpak. On one hand, AI tools can explain and exemplify the use of English idioms. On the other hand, they may lack the data for their Karakalpak entries. In earlier studies, we reported a project-based technology for teaching English food idioms to Karakalpak students using illustrations of English idioms, native expressions for bread (nan), salt (duz), and tea, and English cultural practices. The present paper is a continuation of this research strand, looking at the comparative use of AI tools in teaching English food idioms to Karakalpak learners.

The research objective is to provide theoretical justification and develop a methodological framework for integrating AI tools into the teaching of idiomatic competence of intermediate EFL students. To fulfil the research objective, the following research tasks were identified: to identify the components of idiomatic competence, to examine the roles of different AI tools in teaching idiomatic competence, to design a framework for AI-assisted teaching of idiomatic competence and a typology of tasks, and to suggest criteria and guidance for the critical and responsible use of AI in teaching idiomatic competence. Materials and Methods The workplace of research was the Department of English Linguistics at Karakalpak State University named after Berdakh. The NUKUS · 2026

research work was based on analysis of literature in the fields of idiom teaching, cognitive linguistics, and AI in education; freely available AI tools of different types (conversational chatbots running on large language models, AI-powered writing assistants, text to image generators, speech technologies, quiz generators); and the list of English food idioms taught by the author to intermediate (B1–B2) students X A L Q A R O I L M I Y- A M A L I Y K O N F E R E N S I YA

along with the related idioms in the Karakalpak language. The following methods were used in the study: theoretical analysis and synthesis of literature, functional analysis of AI tools (i.e., each type of tool was considered in terms of what tasks it can be used for in idiom learning), trial queries to AI chatbots regarding meaning, origin, usage, possible Karakalpak equivalents of individual idioms followed by author’s verification of AI output, pedagogical modelling (the stages of AI-assisted instruction were modelled together with relevant tasks), and generalisation of author’s teaching experience. Idiomatic competence has been defined as comprising four interrelated components: a semantic component (understanding of figurative meanings and their « O ‘ Z B E K I S T O N O L I M A AY O L L A R I N I N G M A M L A K AT T A R A Q Q I Y O T I G A Q O ‘ S H AY O T G A N H I S S A S I »

motivation); a pragmatic component (knowledge of situations, interlocutors, and registers in which a given idiom can or cannot be used); a linguo-cultural component (understanding of the cultural background of an idiom and its connections with the student’s native culture); and a productive component (the ability of a student to appropriately use idioms in both speaking and writing). Each of these components has been analysed in relation to the different functions of AI tools. Results 1. Functions of AI tools in idiom learning. The analysis showed that different types of AI tools support different components of idiomatic competence. The results are summarised in Table 1. Table 1. Types of AI tools and their functions in developing idiomatic competence Type of AI tool Function in idiom learning Example of a learning task Conversational chatbots Explaining meaning, origin and Ask the chatbot to explain (large language models) register; generating contextual "to bring home the bacon", examples; role-play dialogue; then write three sentences feedback on learner sentences and request feedback AI writing assistants Detecting unnatural or incorrect Revise a short essay by idiom use; suggesting more replacing literal phrases idiomatic alternatives with suitable food idioms Text-to-image generators Visualising the literal image behind Generate an image of an idiom to support dual coding "walking on eggshells" and explain the gap between literal and figurative meaning

Type of AI tool Function in idiom learning Example of a learning task Speech technologies Modelling pronunciation and stress; Record a dialogue using (text-to-speech, speech checking learners’ spoken five idioms and check it recognition) production with speech recognition NUKUS · 2026

Adaptive quiz and Creating personalised exercises and Generate a gap-fill quiz on flashcard generators spaced repetition idioms the learner answered incorrectly

Conversational chatbots proved to be the most versatile tools. In response to a request to explain an idiom such as to take something with a pinch of salt, they X A L Q A R O I L M I Y- A M A L I Y K O N F E R E N S I YA

typically provide its meaning, a probable origin, information about register and several contextual examples, and they can adapt the explanation to the learner's level on request. They can also act as conversation partners in role-plays in which the learner is asked to use particular idioms, and they can comment on the correctness of the learner's sentences. Text-to-image generators are valuable for the semantic component, since an image of the literal meaning (for example, a person literally walking on eggshells) makes the contrast between literal and figurative meaning vivid and memorable. 2. Limitations identified. Verification of chatbot responses revealed several limitations that are important for teaching. Explanations of the origin of idioms are sometimes presented with more certainty than the historical « O ‘ Z B E K I S T O N O L I M A AY O L L A R I N I N G M A M L A K AT T A R A Q Q I Y O T I G A Q O ‘ S H AY O T G A N H I S S A S I »

evidence allows. When asked for Karakalpak equivalents, chatbots may offer literal translations, expressions from related Turkic languages, or expressions that do not exist in Karakalpak at all, presenting them as established equivalents. Example sentences are generally grammatical, but they sometimes use idioms in an unnatural context or register. These observations confirm that, for a less widely represented language such as Karakalpak, AI output must be treated as a hypothesis to be checked rather than as an authoritative source. 3. A four-stage model of AI-assisted idiom instruction. On the basis of the analysis, a model consisting of four stages was developed: Stage 1 — Discovery. Learners meet idioms in authentic texts or video and try to guess their meaning from context. They then consult an AI chatbot to check their guesses and to learn about the literal image and origin of each idiom. Text-to-image tools are used to visualise the literal meaning.

Stage 2 — Comparison. Learners ask the AI tool whether a Karakalpak equivalent exists and then verify the answer using dictionaries, family members and the teacher. They classify each idiom as having a full equivalent, a partial equivalent with a different image (for example, to bring home the bacon and the Karakalpak expression based on bread), or no equivalent. Discussion of inaccurate AI suggestions becomes a learning opportunity in itself.

Stage 3 — Practice. Learners engage in role-play dialogues with a chatbot, write short texts using the idioms and request feedback, and complete personalised quizzes generated on the basis of their errors. Speech technologies are used to practise pronunciation.

Stage 4 — Creation and reflection. Learners produce their own products — for instance, a dialogue, a short story or an illustrated idiom glossary — in which they use the idioms appropriately. They reflect on what the AI tool did well, where it was wrong, and what they learned by checking it. Conclusions and Recommendations The study shows that AI tools offer significant possibilities for developing all four components of idiomatic competence: they provide immediate, level-appropriate explanations of meaning and NUKUS · 2026

origin, visualise the images underlying idioms, create unlimited opportunities for contextual practice and personalised feedback, and support comparison with the learner's native language. At the same time, their limitations — particularly the risk of inaccurate information about less widely represented languages such as Karakalpak — mean that AI cannot replace the teacher. Its effectiveness depends on X A L Q A R O I L M I Y- A M A L I Y K O N F E R E N S I YA

integration into a structured, teacher-guided methodology. On the basis of the study, the following recommendations are made: (1) introduce AI tools into idiom instruction according to the four-stage model of discovery, comparison, practice and creation; (2) teach students to formulate clear prompts and to request explanations and feedback rather than ready-made texts; (3) systematically verify AI-generated information about native-language equivalents with dictionaries and human informants, and use errors as material for discussion; (4) combine AI tools with collaborative and project-based activities so that « O ‘ Z B E K I S T O N O L I M A AY O L L A R I N I N G M A M L A K AT T A R A Q Q I Y O T I G A Q O ‘ S H AY O T G A N H I S S A S I »

idioms are used in meaningful communication; (5) provide professional development for teachers in the pedagogical use of AI; and (6) establish clear rules on the ethical use of AI in accordance with academic integrity requirements. The next stage of the research will involve the experimental verification of the proposed model with intermediate students of Karakalpak State University, as well as the creation of a verified bank of English–Karakalpak idiom correspondences that could improve the quality of AI-supported learning for Karakalpak learners.