Figure Ilustration AI
FORMOSA NEWS - Yogyakarta - High school students practicing English through artificial intelligence (AI) conversational platforms achieved dramatic breakthroughs in their verbal communication skills, according to a pivotal study published in June 2026. The research, appearing in the Formosa Journal of Sustainable Research, reveals how cutting-edge cloud software transforms traditional foreign language education by turning smartphones and computers into live, judgment-free tutoring assistants. The investigative study was conducted by academic researchers Agussalim Ismail and Ashadib from Universitas Negeri Yogyakarta. Over a multi-month period, the team examined the practical impact of deploying interactive AI conversation partners built within Google AI Studio for eleventh-grade high school classrooms. The findings matter immensely because they establish a scalable, low-cost framework for public schools to overcome chronic shortages of native English instructors while simultaneously dismantling the heavy psychological barriers of language speaking anxiety.
The Challenge of Foreign Language Anxiety
Traditional classroom models consistently struggle to equip non-native students with authentic conversational spontaneity. In typical English as a Foreign Language (EFL) settings, adolescent learners suffer from severe speaking apprehension, driven by a profound fear of peer judgment or making grammatical mistakes in front of an audience. This anxiety leads to minimal active participation and delayed linguistic development. To confront these systemic barriers, the Universitas Negeri Yogyakarta researchers introduced a cloud-based digital solution. The platform operates seamlessly across Windows, macOS, and Android operating systems via web browsers, removing immediate financial or infrastructure hurdles for public school integration.
Simple Framework, Complex Technology
The methodology relied on an explanatory sequential mixed-methods research design. To establish explicit baseline data, the researchers utilized a pre-experimental, one-group pretest-posttest framework involving a full cohort of grade XI students through saturated sampling. The structural flow of the intervention followed a clear trajectory:
Quantifying the Classroom Breakthroughs
Statistical data analysis revealed a profound, uniform rise in oral competence across every single measured linguistic metric ($p < 0.05$). The average holistic student test scores surged from a pre-test baseline of 59.10 up to an impressive post-test average of 72.09. The single largest percentage spike materialized within the fluency domain, which experienced a remarkable 24.1% relative increase. Quantitative data also indicated that pronunciation reached the highest peak score at 73.5 points, driven directly by the software's automated speech recognition capabilities. Beyond the raw scores, student sentiment shifted dramatically. An overwhelming 87% of high school participants reported positive perceptions of the platform, resulting in a strong satisfaction average of 4.20 out of 5.0. Students consistently explained that the AI environment felt safe, comfortable, and entirely non-judgmental.
Implications and Real-World Impact
The success of the platform offers immediate, actionable solutions for global education ministries, structural policymakers, and localized school systems. By offloading repetitive vocal drills, basic vocabulary enforcement, and routine grammar feedback to cloud-based automation tools, institutions can rapidly mitigate regional disparities. This is particularly vital in rural or resource-constrained communities that completely lack access to native English speakers. Crucially, the study clarifies that technology should not displace human professionals. Instead, it advocates for a sophisticated technological-humanistic synergy. While the platform expertly manages scalable oral exercises, human teachers are freed to dedicate their efforts to deep emotional support, cultural immersion, critical thinking development, and nuanced interpersonal communications.
Author Profil
Agussalim Ismail, M.Pd. – Academic researcher and expert in educational technology and English language instruction at Universitas Negeri Yogyakarta.
Ashadib, M.Hum. – Research specialist focused on language acquisition, pedagogical innovations, and applied linguistics at Universitas Negeri Yogyakarta.
Source
The Challenge of Foreign Language Anxiety
Traditional classroom models consistently struggle to equip non-native students with authentic conversational spontaneity. In typical English as a Foreign Language (EFL) settings, adolescent learners suffer from severe speaking apprehension, driven by a profound fear of peer judgment or making grammatical mistakes in front of an audience. This anxiety leads to minimal active participation and delayed linguistic development. To confront these systemic barriers, the Universitas Negeri Yogyakarta researchers introduced a cloud-based digital solution. The platform operates seamlessly across Windows, macOS, and Android operating systems via web browsers, removing immediate financial or infrastructure hurdles for public school integration.
Simple Framework, Complex Technology
The methodology relied on an explanatory sequential mixed-methods research design. To establish explicit baseline data, the researchers utilized a pre-experimental, one-group pretest-posttest framework involving a full cohort of grade XI students through saturated sampling. The structural flow of the intervention followed a clear trajectory:
- Initial Oral Pre-testing: Students completed comprehensive baseline speaking examinations evaluating fluency, pronunciation, grammar, and vocabulary.
- AI Tutor Intervention: Students engaged in adaptive conversation practice using a dedicated application driven by Generative AI and Natural Language Processing (NLP) technologies via Google AI Studio.
- Final Oral Post-testing: Researchers remeasured the identical linguistic criteria following the classroom intervention.
- Qualitative Triangulation: Focus group interviews, semi-structured surveys, and direct classroom observations were integrated with the statistical results to capture student attitudes and behavioral changes.
Quantifying the Classroom Breakthroughs
Statistical data analysis revealed a profound, uniform rise in oral competence across every single measured linguistic metric ($p < 0.05$). The average holistic student test scores surged from a pre-test baseline of 59.10 up to an impressive post-test average of 72.09. The single largest percentage spike materialized within the fluency domain, which experienced a remarkable 24.1% relative increase. Quantitative data also indicated that pronunciation reached the highest peak score at 73.5 points, driven directly by the software's automated speech recognition capabilities. Beyond the raw scores, student sentiment shifted dramatically. An overwhelming 87% of high school participants reported positive perceptions of the platform, resulting in a strong satisfaction average of 4.20 out of 5.0. Students consistently explained that the AI environment felt safe, comfortable, and entirely non-judgmental.
Implications and Real-World Impact
The success of the platform offers immediate, actionable solutions for global education ministries, structural policymakers, and localized school systems. By offloading repetitive vocal drills, basic vocabulary enforcement, and routine grammar feedback to cloud-based automation tools, institutions can rapidly mitigate regional disparities. This is particularly vital in rural or resource-constrained communities that completely lack access to native English speakers. Crucially, the study clarifies that technology should not displace human professionals. Instead, it advocates for a sophisticated technological-humanistic synergy. While the platform expertly manages scalable oral exercises, human teachers are freed to dedicate their efforts to deep emotional support, cultural immersion, critical thinking development, and nuanced interpersonal communications.
Author Profil
Agussalim Ismail, M.Pd. – Academic researcher and expert in educational technology and English language instruction at Universitas Negeri Yogyakarta.
Ashadib, M.Hum. – Research specialist focused on language acquisition, pedagogical innovations, and applied linguistics at Universitas Negeri Yogyakarta.
Source
Agussalim Ismail, Ashadib. AI Tutor-Based English Learning in Improving High School Students' Speaking Skills. Formosa Journal of Sustainable Research (FJSR).Vol. 5, No. 6, Halaman 409-428
DOI:https://doi.org/10.55927/fjsr.v5i6.40
URL:https://journalfjsr.my.id/index.php/fjsr
DOI:
URL:
0 Komentar