INTERNATIONAL MULTIDISCIPLINARY JOURNAL FOR
RESEARCH & DEVELOPMENT
SJIF 2019: 5.222 2020: 5.552 2021: 5.637 2022:5.479 2023:6.563 2024: 7,805
eISSN :2394-6334 https://www.ijmrd.in/index.php/imjrd Volume 12, issue 07 (2025)
69
AI-ENHANCED PHYSICS INSTRUCTION IN UZBEK: EVALUATING
COMPREHENSION OF SCIENTIFIC TERMINOLOGY USING LANGUAGE
MODELS
Navbahor Qurbanbayeva Shermat kizi
Berdaq Karakalpak State University,
Faculty of Physics, Department of Physics
Abstract:
This study explores the use of artificial intelligence-based tools, particularly large
language models (LLMs), in teaching physics terminology in the Uzbek language. The research
focuses on the effectiveness of AI-assisted instruction in improving students’ understanding of
core physics concepts expressed in Uzbek. A comparative experimental methodology was
employed: one group received traditional instruction, while the other engaged with interactive,
AI-supported lessons using localized terminology. The outcomes were evaluated through
comprehension tests, semantic accuracy checks, and student feedback. The results indicate that
the AI-driven approach significantly enhances learners’ grasp of scientific terms, promotes
linguistic clarity, and fosters deeper conceptual understanding in native-language physics
education.
Keywords:
Uzbek language, physics education, scientific terminology, artificial intelligence,
large language models, AI in education, native-language instruction, comprehension assessment,
interactive learning, AI-assisted teaching.
In recent years, artificial intelligence (AI), and more specifically large language models (LLMs),
have emerged as powerful tools in transforming educational landscapes. While much of the
existing research on AI-assisted learning has focused on English-medium instruction, less
attention has been given to AI applications in native-language science education—particularly
in languages such as Uzbek. Physics, being one of the most terminology-intensive disciplines,
poses a unique challenge for students when taught in their native language, especially when
scientific terms are translated or adapted from global standards.
This study addresses that gap by investigating how AI can be utilized to improve the
understanding of physics terminology in Uzbek. We hypothesize that AI-based educational
tools, when localized linguistically and culturally, can support more effective comprehension
and retention of scientific concepts. By integrating AI into lesson delivery, and designing tasks
that promote semantic understanding, this research seeks to evaluate how students interact with,
absorb, and retain physics terminology presented in their mother tongue.
Through a controlled experimental design, the study compares traditional instruction methods
with AI-supported lessons, using ChatGPT and similar models fine-tuned or prompted for
Uzbek terminology explanation. The goal is to determine whether AI can serve not only as a
linguistic assistant but also as a pedagogical partner in deepening students' conceptual and
terminological knowledge in physics.
This research employed a quasi-experimental design with two student groups from a secondary
school physics program in Uzbekistan. The control group was taught using conventional
teaching methods, including lectures, textbook-based explanation, and chalkboard problem-
INTERNATIONAL MULTIDISCIPLINARY JOURNAL FOR
RESEARCH & DEVELOPMENT
SJIF 2019: 5.222 2020: 5.552 2021: 5.637 2022:5.479 2023:6.563 2024: 7,805
eISSN :2394-6334 https://www.ijmrd.in/index.php/imjrd Volume 12, issue 07 (2025)
70
solving. The experimental group, by contrast, received AI-supported instruction using ChatGPT
and other Uzbek-compatible language models.
The AI tools were prompted in Uzbek to explain core physics terms such as
harorat
(temperature),
zarralar
(particles),
elektr toki
(electric current), and
kinetik energiya
(kinetic
energy) in simple, accessible language. Instructional content was delivered in interactive Q&A
formats, where students posed questions to the AI in Uzbek and received real-time responses.
These interactions were moderated by the teacher to ensure relevance and accuracy.
To evaluate comprehension, both groups completed standardized tests designed to measure
three key domains: (1) terminological recognition, (2) conceptual understanding, and (3)
contextual usage of physics terms. In addition, qualitative data were collected through
structured interviews and surveys that explored students’ perceptions of clarity, confidence, and
engagement during the learning process.
The results revealed a statistically significant improvement in the experimental group’s ability
to comprehend and use physics terminology in Uzbek. On average, the AI-supported group
scored 23% higher on conceptual understanding questions and 19% higher on terminology
recognition tasks compared to the control group.
Students in the AI-assisted group demonstrated greater semantic clarity when defining and
applying scientific terms in written and oral responses. For example, over 80% of the
experimental group correctly described the difference between
issiqlik
(heat) and
harorat
(temperature), compared to only 52% in the control group.
Qualitative feedback highlighted increased student engagement and curiosity. Many students
noted that interacting with AI in their native language made them feel more confident and less
intimidated by abstract physics concepts. Some also mentioned the benefit of being able to ask
follow-up questions without time pressure or judgment.
The findings suggest that AI-powered instruction—when localized in the Uzbek language—can
significantly improve comprehension of scientific terminology. By serving as a responsive,
always-available learning assistant, AI tools help bridge the gap between abstract scientific
language and native linguistic intuitions.
This method appears particularly effective in enhancing physics education in multilingual
contexts, where students may face dual challenges: understanding the science itself and
interpreting unfamiliar terminology. AI tools like ChatGPT enable teachers to offer
individualized explanations at scale, addressing different levels of prior knowledge.
Nonetheless, the study also revealed several challenges. Some students tended to over-rely on
AI without critically analyzing the responses. In addition, the accuracy of AI-generated Uzbek
explanations varied depending on prompt quality and model limitations. These observations
point to the need for teacher facilitation and careful prompt engineering in AI-assisted
classrooms.
This study demonstrates that artificial intelligence, specifically large language models like
ChatGPT, can play a valuable role in helping students understand complex physics terminology
INTERNATIONAL MULTIDISCIPLINARY JOURNAL FOR
RESEARCH & DEVELOPMENT
SJIF 2019: 5.222 2020: 5.552 2021: 5.637 2022:5.479 2023:6.563 2024: 7,805
eISSN :2394-6334 https://www.ijmrd.in/index.php/imjrd Volume 12, issue 07 (2025)
71
in the Uzbek language. AI-supported lessons enhanced conceptual clarity, increased
engagement, and led to measurable learning gains.
To ensure successful integration of AI in physics instruction, educators should combine
traditional pedagogy with AI’s interactivity and linguistic adaptability. Future research may
explore the use of fine-tuned Uzbek language models and longitudinal impacts of AI-supported
science education.
The positive outcomes of this study suggest a strong potential for AI as a scalable, inclusive,
and linguistically sensitive tool for teaching STEM subjects in underrepresented languages.
References
1.
OpenAI. (2023).
GPT-4 Technical Report
.
2.
UNESCO. (2021).
Artificial Intelligence and the Futures of Education: Challenges and
Opportunities
.
3.
Wang, Y., & Blei, D. (2020). “Language Models for Science Education.”
IEEE
Transactions on Learning Technologies
.
4.
Tashkenbaev, B. (2022).
Fizik atamalar va ularning o‘zbek tilidagi muqobillari
.
Toshkent: Fan nashriyoti.
5.
Mitrovic, A., et al. (2022). “AI in STEM Classrooms: Potentials and Pitfalls.”
Educational Technology Research & Development
.
6.
Ravshanov, R. (2023). “Sun’iy intellekt yordamida ta’lim samaradorligini oshirish.”
O‘zbekiston Pedagogik Jurnali
, 1(3), 18–25.
7.
Brown, T. et al. (2020). “Language Models Are Few-Shot Learners.”
Advances in
Neural Information Processing Systems (NeurIPS)
.
8.
Amershi, S., et al. (2019). “Guidelines for Human-AI Interaction.”
Proceedings of the
ACM CHI Conference on Human Factors in Computing Systems
.
9.
Salakhov, A. & Karimov, D. (2023). “Native Language Support in AI-Based Physics
Instruction.”
International Journal of Educational Technology in Higher Education
.
10.
Rahimova, M. (2023).
Didaktik yondashuvlarda sun’iy intellekt texnologiyalari
. Termiz:
TDPU Nashriyoti.
