MODERN EDUCATION AND DEVELOPMENT
Выпуск журнала №-28
Часть–1_Июнь –2025
302
AI-POWERED DIGITAL TOOLS FOR ENHANCING SPEAKING
PROFICIENCY FOR FOREIGN LANGUAGE STUDENTS
Umarov Ozodbek
Abstract: The research explores how artificial intelligence powered digital
devices affect the speaking proficiency acquisition of foreign language students.
Language acquisition becomes more achievable through the implementation of
artificial intelligence (AI) since it provides personalized and accessible and
interactive speaking practice for students. The study used a mixed research design
that measured 60 English as a Foreign Language (EFL) university students who
utilized speech recognition applications and chatbots together with voice analysis
tools for a 5-week duration. The research used Common European Framework of
Reference (CEFR)-based pre- and post-tests to evaluate changes in speaking ability.
Learners participated in semi-structured interview sessions for exploring their
insights about their learning experiences. Participants achieved statistically
important advancement in their overall speaking skill levels mainly through improved
pronunciation and fluency results. The tools connected to artificial intelligence
provided participants with enhanced confidence levels together with heightened
motivational states and student engagement because of their flexible function and
quick feedback mechanisms. In addition to their noted strengths the tools displayed
some weaknesses through their repetitive nature and occasional inaccuracies in
feedback. The research identifies AI digital instruments as valuable supports for
conventional speech learning because they promote student autonomy through
continuous practice. Further studies should examine both prolonged effects of these
tools and ways to enhance their production methods.
Key words: Artificial Intelligence (AI), Speaking Proficiency, Speech
Recognition, Chatbots, Voice Analysis Tools, CEFR (Common European Framework
of Reference),Real-time Feedback, Mobile-Assisted Language Learning (MALL),
Autonomous Learning, Anxiety Reduction, Mixed Methods, Speaking Assessment
MODERN EDUCATION AND DEVELOPMENT
Выпуск журнала №-28
Часть–1_Июнь –2025
303
Introduction
Artificial intelligence enabled educational approaches to develop new
prospects by creating interactive learning systems. The acquisition of foreign
languages creates difficulties for students to get live feedback because they struggle
with speaking without anxiety and rarely engage in authentic language usage.
Education occurs successfully through traditional teaching practices despite their
insufficient provision of real-time guidance for each student and limited practice time
away from material content. The language learning field experiences transformations
because of artificial intelligence-based digital instruments which include speech
recognition systems as well as virtual partners and pronunciation assessment tools.
Educational artificial intelligence tools enable students to receive real-time feedback
during virtual conversations and feedback monitoring which creates better
opportunities for developing independent speaking abilities. All students in this
generation have effortless access to technology through personal computers and
smartphones because they function both inside educational establishments and in
public spaces. Future academic research must evaluate the precise influence of
technological tools on student speaking development because the technology market
expands without adequate proof. The study analyzes AI digital tool effects on
language speaking abilities by uniting research from technological practice and
pedagogical strategies and speech development frameworks. The assessment
measures examine benefits and drawbacks which will assist faculty members and
developers together with learners to enhance their foreign language speaking abilities.
Methodology
The research used mixed methods to identify how artificial intelligence
software tools help foreign language students develop their speaking abilities.
Different types of data collection incorporating quantitative and qualitative research
methods allowed the investigation to provide deep insights about learner's
improvement and their educational experiences.
Participants
MODERN EDUCATION AND DEVELOPMENT
Выпуск журнала №-28
Часть–1_Июнь –2025
304
Sixty students studying intermediate-level foreign language English as a
Foreign Language courses in two universities participated in the study. The research
team chose participants through purposive sampling because all recruits demonstrated
basic digital skills and possession of smartphones or computers. This research
evaluation included three AI-assisted speaking programs: (1) a speech identification
application that delivers real-time pronunciation evaluation, (2) a language-learner
specialized virtual dialogue chatbot and (3) a voice analytic system which monitors
speech precision and speed. The speaking improvement was evaluated through pre-
test and post-test assessments which followed speaking descriptors from the Common
European Framework of Reference (CEFR). Research investigators utilized semi-
structured protocols during interviews to understand the learners’ perspectives about
their interactions with the supplied tools. The entire research period lasted six weeks.
The participants performed the speaking pre-test during the initial week of the study.
The participants spent twenty minutes daily using AI instruments for five weeks
during their speaking practice. The participants completed post-testing during the last
week while also taking part in interview sessions. Researchers recorded all speaking
activities and communication for subsequent analytical purposes. Paired sample t-
tests were used to analyze pre- and post-test quantitative data in order to measure any
significant change in speaking proficiency. Researchers applied thematic coding to
interview data to extract recurring patterns about student perceptions of working with
AI tools.
Results
The research findings based on quantitative and qualitative analysis showed
that students achieved better speaking ability results from continuously using AI
digital tools over five weeks. The speaking proficiency scores increased significantly
according to the CEFR-based speaking rubric assessments. The pre-test scores began
at 5.2 but post-test scores improved to 6.1 across the board thus demonstrating
noteworthy progress. Researchers applied a paired sample t-test and obtained results
of t(59) = 7.89 that demonstrated statistical significance (p < 0.01) to verify the mean
difference. The AI speech recognition app delivered 0.9-point pronunciation
MODERN EDUCATION AND DEVELOPMENT
Выпуск журнала №-28
Часть–1_Июнь –2025
305
enhancements to students through multiple uses because students demonstrated
reduced phonetic problems. Students scored 0.8 points higher on fluency tests because
they engaged in multiple relaxed conversations with the AI assistant. Students who
utilized the voice analysis tool featuring filler word tracking and sentence structure
analysis showed moderate improvement in both range of grammar and vocabulary
(+0.5 points). The improved results show that AI-based tools created helpful feedback
structures to guide students during their independent speaking exercises. Speaking
through an AI interface appeared more comfortable to learners than speaking with
human teachers or their peers. The lower levels of anxiety among learners enabled
them to expand their speaking sessions duration and frequency. Learners said they
would continue with another attempt after getting immediate feedback containing
positive reinforcement and corrective comments. The participants valued receiving
detailed feedback instantly from the system particularly regarding pronunciation
along with intonation. The students developed better recognition of their speech
patterns which caused them to naturally make self-corrections during speaking
practice. The AI tools faced sporadic problems with pronunciation scoring according
to some participants amongst the group. Ease of schedule flexibility made practicing
at any time during or outside their normal classes very accessible for students.
Students took advantage of the tools during their daily commute and short
breaks as a strategy to integrate speaking practice into their daily schedule. Students
used these accessible tools with regularity because they could access them at any time.
The participants liked the chatbot on average yet they found difficulty in interactions
because its conversations often repeated the same lines and sounded too rigid. At the
beginning of their use the digital literacy skills of some learners prevented them from
effectively using mobile applications. The research findings show that AI digital tools
effectively enhanced student speaking ability according to both the numerical and
textual data analysis. Test scores improved together with learner mental attitudes and
increased motivation and speaking comfort.
Discussion
MODERN EDUCATION AND DEVELOPMENT
Выпуск журнала №-28
Часть–1_Июнь –2025
306
The study proves that AI digital tools effectively boost foreign language
speaking abilities of learners through personalized learning opportunities which are
both accessible and engaging. The observed advancements in CEFR speaking scores
together with learner appreciation indicate that these technological tools deliver
effective results that students discreetly welcome. The main result from this research
showed students became better at both pronunciation and fluency. Research from Li
et al. (2020) confirmed that real-time corrective feedback remains essential for
developing speaking competencies because students obtained improved speaking
scores. Students who used speech recognition technology became better at
recognizing pronunciation mistakes while using the chatbot they produced better
sustained oral responses which helped their fluency increase. The findings support
second language acquisition models which state spoken competence development
needs both output from learners and feedback and repetition (Swain, 1985; Ellis,
2003).
The gathered qualitative findings showed learners achieved elevated levels of
self-confidence alongside higher motivation since these elements remain vital to
language acquisition. The low-stress environment alongside uncontested interactions
served to lessen the usual fear which students experience while communicating in
foreign languages. Students who avoid speaking in classroom environments because
of fear or concern about mistakes need this approach to learn effectively. The
accessibility together with flexible nature of these tools enabled students to integrate
their speaking practice into their regular daily activities. The concept that mobile-
assisted language learning (MALL) generates better and substantive language
encounters outside classroom walls (Stockwell, 2013) finds support. While AI tools
deliver numerous benefits they are not sufficient replacements for human
communication or curriculum-developed instruction because students experience
repeated dialogue cycles and technical system problems.
Conclusion
A research project evaluated how artificial intelligence tools help develop
speaking abilities for language students with emphasis on pronunciation and learner
MODERN EDUCATION AND DEVELOPMENT
Выпуск журнала №-28
Часть–1_Июнь –2025
307
fluency along with participation levels. Participants devoted five weeks of their
studies to AI applications consisting of speech recognition systems and virtual
chatbots together with voice analysis tools. The research based on mixed data samples
revealed strong positive outcomes for trainees' capability in oral communication. The
implemented AI applications led to statistically confirmed improvement of speaking
test assessments according to CEFR standards where pronunciation received notable
gains alongside fluent performances. The data indicates that AI tools provide
solutions to the common barriers which prevent foreign language speaking
development through insufficient real-time feedback and small amounts of practice.
The implementation of real-time error detection combined with repeated practice
along with performance measurement systems apparently builds self-governing and
analytical speaking practices among students. Additional information gained through
interview data provided deeper knowledge about the learners' experiences. Several
interviewees praised the tools because they found them both interesting and beneficial
for building confidence and suitable for regular everyday use. The AI-generated non-
judgmental communication approach led learners to practice more often because it
reduced their anxiety which matches well with affective practices in language
acquisition. The learners valued unlimited practices that accommodated unrestricted
speaking time because this enabled them to improve their rhythm and receive
confidence in their speech together with enhanced stamina. Additional constraints in
AI speaking tool technology were found during the conducted research. The
participants mentioned the constrained nature of chatbot chat because conversations
lacked contextual depth which could restrict the growth of more complex
conversational abilities. Particular speech instances and scoring fluctuations that
occurred were recorded in the study which might reduce the reliability factor when
delivering feedback during some interactions. The current implementation of AI
speaking tools requires continuous development because they must achieve alignment
between educational requirements and student reception standards.
MODERN EDUCATION AND DEVELOPMENT
Выпуск журнала №-28
Часть–1_Июнь –2025
308
REFERENCES
1.
Ali, Z., & Ahmad, S. (2022). Artificial intelligence in language learning: A
systematic review of tools and trends. Journal of Educational Technology & Online
Learning, 5(1), 22–35. https://doi.org/10.1016/j.edtech.2022.01.003
2.
Chen, X., Zou, D., & Xie, H. (2021). Fostering oral language learning with AI-
based chatbots: A systematic review of recent research. Computer Assisted Language
Learning, 34(8), 1013–1039. https://doi.org/10.1080/09588221.2020.1814647
3.
Ellis, R. (2003). Task-based language learning and teaching. Oxford University
Press.
4.
Li, V., Chau, C. H., & Wong, B. T. M. (2020). Enhancing speaking skills with
AI speech recognition: The learners’ perceptions. Language Learning & Technology,
24(2), 123–142.
5.
Lu, X., & Ai, H. (2015). Automatic pronunciation error detection in second
language
learning:
A
review.
Speech
Communication,
72,
1–15.
https://doi.org/10.1016/j.specom.2015.04.001
6.
Stockwell, G. (2013). Mobile-assisted language learning: Concepts, contexts,
and challenges. In M. Thomas, H. Reinders & M. Warschauer (Eds.), Contemporary
computer-assisted language learning (pp. 201–216). Bloomsbury.
7.
Swain, M. (1985). Communicative competence: Some roles of comprehensible
input and comprehensible output in its development. In S. Gass & C. Madden (Eds.),
Input in second language acquisition (pp. 235–253). Newbury House.
8.
Wang, Y., & Vasquez, C. (2012). Web 2.0 and second language learning: What
does the research tell us? CALICO Journal, 29(3), 412–430.
9.
Zhang, Y., & Zou, D. (2022). A review of AI applications in language
education: Benefits, challenges, and future directions. ReCALL, 34(1), 1–18.
https://doi.org/10.1017/S0958344021000162
