ENHANCING EFL SPEAKING THROUGH AI
VOICE INTERACTION: A QUALITATIVE CASE
STUDY OF A LEARNER WITH STARGARDT
DISEASE.
MEJORA DE LA EXPRESIÓN ORAL EN EFL MEDIANTE LA
INTERACCION POR VOZ CON IA: UN ESTUDIO DE CASO
CUALITATIVO DE UN ESTUDIANTE CON ENFERMEDAD DE
STARGARDT.
Arianna Stefania Peralta Cedeño
Universidad Laica Eloy Alfaro de ManabÃ
Eder Intriago Palacios
Universidad Laica Eloy Alfaro de ManabÃ
pág. 4487
DOI: https://doi.org/10.37811/cl_rcm.v10i4.25451
Enhancing EFL speaking through AI voice interaction: A qualitative case
study of a learner with Stargardt disease.
Arianna Stefania Peralta Cedeño1
peraltaariannast@gmail.com
https://orcid.org/0009-0000-8545-9539
Universidad Laica Eloy Alfaro de ManabÃ
Ecuador
Eder Intriago Palacios
eder.intriago@uleam.edu.ec
https://orcid.org/0000-0002-9433-7186
Universidad Laica Eloy Alfaro de ManabÃ
Ecuador
ABSTRACT
Artificial Intelligencediagnosed voice asistants have emerged as valuable tools for supporting English
as a Foreign Language (EFL) Learning by providing accesible, personalized, and low anxiety
opportunities for speaking practice. However, limited research has examined their roles as assistive
technologies for learners with visual impairments. This qualitative instrumental case study explored how
Gemini’s voice assisstant supported the speaking development, confidence, and autonomy of an adult
EFL learner diagnosed with Stargardt disease. Data were collected through a semi-structured interviw
and analyzed usign thematic analysis. Five themes emerged: accessibility as linguistic empowerment,
AI mediated speaking practice, autonomy and metacognitive regulation, reduced speaking anxiety, and
perceived limitations of AI-assisted learning. The findings indicate that AI voice assistant functioned
both, as a language learning partner and as an accessibility tool, enabling authentic oral practice while
reducing dependence on text-based interfaces. The participant reported grater confidence, increased
willingness to communicate, and improved autonomy through regular voice interaction. Despite
limitations, incljuding the lack of personalized process tfracking and the inability to replace authentic
human interacti9on, the study hoghlightd the potential of AI voicfe technology to foster inclusive
language learning and support speaking development for learners with visual impairments.
Keywords: artificial intelligence; voice assistants; EFL speaking; Stargardt; English.
1 Autor principal
Correspondencia: peraltaariannast@gmail.com
DOI: https://doi.org/10.37811/cl_rcm.v10i4.25451
Enhancing EFL speaking through AI voice interaction: A qualitative case
study of a learner with Stargardt disease.
Arianna Stefania Peralta Cedeño1
peraltaariannast@gmail.com
https://orcid.org/0009-0000-8545-9539
Universidad Laica Eloy Alfaro de ManabÃ
Ecuador
Eder Intriago Palacios
eder.intriago@uleam.edu.ec
https://orcid.org/0000-0002-9433-7186
Universidad Laica Eloy Alfaro de ManabÃ
Ecuador
ABSTRACT
Artificial Intelligencediagnosed voice asistants have emerged as valuable tools for supporting English
as a Foreign Language (EFL) Learning by providing accesible, personalized, and low anxiety
opportunities for speaking practice. However, limited research has examined their roles as assistive
technologies for learners with visual impairments. This qualitative instrumental case study explored how
Gemini’s voice assisstant supported the speaking development, confidence, and autonomy of an adult
EFL learner diagnosed with Stargardt disease. Data were collected through a semi-structured interviw
and analyzed usign thematic analysis. Five themes emerged: accessibility as linguistic empowerment,
AI mediated speaking practice, autonomy and metacognitive regulation, reduced speaking anxiety, and
perceived limitations of AI-assisted learning. The findings indicate that AI voice assistant functioned
both, as a language learning partner and as an accessibility tool, enabling authentic oral practice while
reducing dependence on text-based interfaces. The participant reported grater confidence, increased
willingness to communicate, and improved autonomy through regular voice interaction. Despite
limitations, incljuding the lack of personalized process tfracking and the inability to replace authentic
human interacti9on, the study hoghlightd the potential of AI voicfe technology to foster inclusive
language learning and support speaking development for learners with visual impairments.
Keywords: artificial intelligence; voice assistants; EFL speaking; Stargardt; English.
1 Autor principal
Correspondencia: peraltaariannast@gmail.com
pág. 4488
Mejora de la expresión oral en EFL mediante la interaccion por voz con IA:
un estudio de caso cualitativo de un estudiante con enfermedad de Stargardt.
RESUMEN
La intligencia artificial ofrece nuevas posibilidades para apoyar el aprendizaje del inglés como lengua
extranjera, especialmente en la práctica de la expresión oral. Sin embargo, su potencial como herramieta
de accesibilidad para estudiantes con discapacidad visual aun ha sido poco explorado. Este esztudio de
caso cualitativo de carácter instrumental tuvo como objetivo explorar como la interaccion por voz
mediante IA puede contribuir al desarrollo de la expresión oral, al confianza y la autonomia de una
estudiante con enfermedad de Stargardt, un transtorno degenerativo de la retina que afecta
principalmente la cision central. Los datos se recopilaron mediante una entrevista semiestructurada y se
analizaron a través del análisis temático, Los hallazgos permitieron identificar cinco temas principales:
accesibilidad y empoderamiento linguistico, practica oralmediada por Inteligencia Artifial, autonomia y
autoregulación metacognitiva, reduccion de la ansiedad al hablar y limitaciones del aprendizaje asistido
por IA. Los resultados sugieren que el asistente de voz funcionó tanto como recurso de aprendizaje como
herramienta de accesibilidad, facilitando la práctica del inglés en situaciones cotidianas y reduciendo la
dependencia de interfaces basadas en texto. Asimismo, la participante reportó mayor confianza,
disposición para comunicarse y autonomÃa, aunque señaló limitaciones relacionadas con el seguimiento
personalizado del progreso y la interacción humana.
Palabras clave: Inteligencia Artificial; expresion oral; asistente de voz; Stargardt; inglés.
ArtÃculo recibido 20 mayo 2026
Aceptado para publicación: 20 junio 2026
Mejora de la expresión oral en EFL mediante la interaccion por voz con IA:
un estudio de caso cualitativo de un estudiante con enfermedad de Stargardt.
RESUMEN
La intligencia artificial ofrece nuevas posibilidades para apoyar el aprendizaje del inglés como lengua
extranjera, especialmente en la práctica de la expresión oral. Sin embargo, su potencial como herramieta
de accesibilidad para estudiantes con discapacidad visual aun ha sido poco explorado. Este esztudio de
caso cualitativo de carácter instrumental tuvo como objetivo explorar como la interaccion por voz
mediante IA puede contribuir al desarrollo de la expresión oral, al confianza y la autonomia de una
estudiante con enfermedad de Stargardt, un transtorno degenerativo de la retina que afecta
principalmente la cision central. Los datos se recopilaron mediante una entrevista semiestructurada y se
analizaron a través del análisis temático, Los hallazgos permitieron identificar cinco temas principales:
accesibilidad y empoderamiento linguistico, practica oralmediada por Inteligencia Artifial, autonomia y
autoregulación metacognitiva, reduccion de la ansiedad al hablar y limitaciones del aprendizaje asistido
por IA. Los resultados sugieren que el asistente de voz funcionó tanto como recurso de aprendizaje como
herramienta de accesibilidad, facilitando la práctica del inglés en situaciones cotidianas y reduciendo la
dependencia de interfaces basadas en texto. Asimismo, la participante reportó mayor confianza,
disposición para comunicarse y autonomÃa, aunque señaló limitaciones relacionadas con el seguimiento
personalizado del progreso y la interacción humana.
Palabras clave: Inteligencia Artificial; expresion oral; asistente de voz; Stargardt; inglés.
ArtÃculo recibido 20 mayo 2026
Aceptado para publicación: 20 junio 2026
pág. 4489
INTRODUCTION
Developing speaking skills is one of the most important yet challenging goals in English as a Foreign
Language education since many learners struggle to express themselves fluently and confidently, even
after years of study. Most EFL learners experience speaking anxiety which is related to the fear of
negative evaluation (Aydin, 2008), making mistakes and peer evaluation. This anxiety limits their
willingness to communicate effectively in English. In addition, learners face an extra challenge: limited
exposure to L2 outside of the classroom, which reduces opportunities for authentic practice and hinders
development of communicative competence. As a result of all this background issues, improving oral
proficiency and build learners’ confidence have become a central common concern among EFL teachers.
In recent years, the rapid advancement of Artificial Intelligence (AI) technologies has opened new
possibilities for enhancing language learning experiences. Among the most popular there are ChatGPT,
Gemini, Perplexity, Stimuler and various others. Each one is integrated with a voice assistant feature
which has gained attention for their potential to provide interactive, personalized and low-anxiety
speaking practice environments.
Previous research indicates that AI chatbots and its voice assistant show promise for enhancing EFL
learners’ speaking skills and confidence. For instance, Qodirqulova (2025) study reports improvements
in fluency, pronunciation accuracy, and increased learner willingness to communicate after chatbot
interventions. Moreover, in the study conducted by Maysuroh et.al (2025) learners perceive chatbots as
low-anxiety, accessible speaking partners that support motivation and autonomy. However, Belda and
Calvo (2022) note limitations such as lack of contextual awareness in some chatbots and the need for
pedagogical scaffolding.
While existing studies have examined AI chatbots in classroom contexts, less attention has been given
to how AI voice interaction may function as an assistive tool for learners with visual impairments. For
individuals experiencing degenerative visual conditions such as Stargardt disease; a genetic disorder
that progressively affects central vision, voice-based technologies may represent not only a language
learning resource but also an accessible communicative support system. In such cases, AI voice
assistants may reduce barriers associates with text-based interaction and create opportunities for
autonomous oral practice.
INTRODUCTION
Developing speaking skills is one of the most important yet challenging goals in English as a Foreign
Language education since many learners struggle to express themselves fluently and confidently, even
after years of study. Most EFL learners experience speaking anxiety which is related to the fear of
negative evaluation (Aydin, 2008), making mistakes and peer evaluation. This anxiety limits their
willingness to communicate effectively in English. In addition, learners face an extra challenge: limited
exposure to L2 outside of the classroom, which reduces opportunities for authentic practice and hinders
development of communicative competence. As a result of all this background issues, improving oral
proficiency and build learners’ confidence have become a central common concern among EFL teachers.
In recent years, the rapid advancement of Artificial Intelligence (AI) technologies has opened new
possibilities for enhancing language learning experiences. Among the most popular there are ChatGPT,
Gemini, Perplexity, Stimuler and various others. Each one is integrated with a voice assistant feature
which has gained attention for their potential to provide interactive, personalized and low-anxiety
speaking practice environments.
Previous research indicates that AI chatbots and its voice assistant show promise for enhancing EFL
learners’ speaking skills and confidence. For instance, Qodirqulova (2025) study reports improvements
in fluency, pronunciation accuracy, and increased learner willingness to communicate after chatbot
interventions. Moreover, in the study conducted by Maysuroh et.al (2025) learners perceive chatbots as
low-anxiety, accessible speaking partners that support motivation and autonomy. However, Belda and
Calvo (2022) note limitations such as lack of contextual awareness in some chatbots and the need for
pedagogical scaffolding.
While existing studies have examined AI chatbots in classroom contexts, less attention has been given
to how AI voice interaction may function as an assistive tool for learners with visual impairments. For
individuals experiencing degenerative visual conditions such as Stargardt disease; a genetic disorder
that progressively affects central vision, voice-based technologies may represent not only a language
learning resource but also an accessible communicative support system. In such cases, AI voice
assistants may reduce barriers associates with text-based interaction and create opportunities for
autonomous oral practice.
pág. 4490
Speaking is broadly recognized as one of the most challenging skills for English students to master,
especially in contexts where it is taught as a foreign language (Bygate, 2018). Learners tend to struggle
with fluency, vocabulary, pronunciation, and grammatical accuracy. (Goh & Burns, 2002). This can be
attributed to the fact that learners often face limited exposure of the target language (Huang, 2024).
Moreover, students frequently lack meaningful opportunities to use English in authentic communicative
situations, which hinders their ability to develop oral proficiency (Richards, 2008). Additionally,
psychological factors such as anxiety and low-self- confidence can significantly affect learners’
willingness to speak during classes (Tuan & Mai, 2015). This study adopts the lens of AI Chatbots with
Voice Assistants, which empowers learners towards independent learning, offering tools that caters to
individual learning paces (Huang, 2024). As several studies have demonstrated that the use of Artificial
Intelligence addresses the psychological barrier to speaking by offering a safe and engaging learning
environment (Busso & Sanchez 2024), reducing speaking anxiety, and creating a feeling of satisfaction
and confidence (Celik et.al 2025).
Artificial intelligence and voice assistants
The integration of artificial intelligence (AI) and voice technologies has significantly reshaped language
education, offering learners more interactive and personalized learning experiences. Busso and Sanchez
(2024) argue that AI tools foster communicative competence by simulating authentic interactional
contexts, allowing learners to engage in real-time dialogue. Similarly, Zhou (2023) emphasizes that AI-
driven voice assistants facilitate oral practice and constructive communication, helping learners enhance
fluency and spontaneity. Research by Kanoksilapatham and Takrudkaew (2025) highlights the
pedagogical importance of integrating ChatGPT to promote communication skills, noting that effective
use depends on careful instructional design. In line with this, PondelÃková and Luprichová (2025) found
that AI chatbots support specialized communication, particularly in academic and professional English
contexts. However, Poaquiza and EstefanÃa (2024) caution that while AI-based voice chatbots can
improve pronunciation and interactional accuracy, they must be complemented with human feedback to
ensure authentic learning outcomes. Collectively, these studies underscore the transformative role of AI
and voice technologies in language learning. Yet, despite their proven utility, limited research has
examined how adult EFL learners emotionally experience and adapt to AI-mediated oral
Speaking is broadly recognized as one of the most challenging skills for English students to master,
especially in contexts where it is taught as a foreign language (Bygate, 2018). Learners tend to struggle
with fluency, vocabulary, pronunciation, and grammatical accuracy. (Goh & Burns, 2002). This can be
attributed to the fact that learners often face limited exposure of the target language (Huang, 2024).
Moreover, students frequently lack meaningful opportunities to use English in authentic communicative
situations, which hinders their ability to develop oral proficiency (Richards, 2008). Additionally,
psychological factors such as anxiety and low-self- confidence can significantly affect learners’
willingness to speak during classes (Tuan & Mai, 2015). This study adopts the lens of AI Chatbots with
Voice Assistants, which empowers learners towards independent learning, offering tools that caters to
individual learning paces (Huang, 2024). As several studies have demonstrated that the use of Artificial
Intelligence addresses the psychological barrier to speaking by offering a safe and engaging learning
environment (Busso & Sanchez 2024), reducing speaking anxiety, and creating a feeling of satisfaction
and confidence (Celik et.al 2025).
Artificial intelligence and voice assistants
The integration of artificial intelligence (AI) and voice technologies has significantly reshaped language
education, offering learners more interactive and personalized learning experiences. Busso and Sanchez
(2024) argue that AI tools foster communicative competence by simulating authentic interactional
contexts, allowing learners to engage in real-time dialogue. Similarly, Zhou (2023) emphasizes that AI-
driven voice assistants facilitate oral practice and constructive communication, helping learners enhance
fluency and spontaneity. Research by Kanoksilapatham and Takrudkaew (2025) highlights the
pedagogical importance of integrating ChatGPT to promote communication skills, noting that effective
use depends on careful instructional design. In line with this, PondelÃková and Luprichová (2025) found
that AI chatbots support specialized communication, particularly in academic and professional English
contexts. However, Poaquiza and EstefanÃa (2024) caution that while AI-based voice chatbots can
improve pronunciation and interactional accuracy, they must be complemented with human feedback to
ensure authentic learning outcomes. Collectively, these studies underscore the transformative role of AI
and voice technologies in language learning. Yet, despite their proven utility, limited research has
examined how adult EFL learners emotionally experience and adapt to AI-mediated oral
pág. 4491
communication, indicating a need for more context-specific exploration.
EFL and speaking development
Recent research confirms that AI-based applications, particularly ChatGPT, have emerged as promising
tools to enhance EFL learners’ speaking abilities. Celik et al. (2025) report that ChatGPT functions as
an effective virtual speaking tutor, significantly increasing learners’ self-efficacy and motivation to
speak. Similarly, Yildiz (2024) demonstrates that consistent ChatGPT use fosters greater linguistic
confidence and autonomy, especially when integrated into blended learning environments. Nuñez et al.
(2025) also observed measurable improvement in learners’ pronunciation, fluency, and coherence when
using ChatGPT’s voice function, emphasizing its potential for individualized oral feedback.
Complementing these findings, Huang (2024) shows that AI voice prompts can provide detailed, real-
time speaking feedback that teachers often cannot deliver due to time constraints. Nonetheless,
Kanoksilapatham and Takrudkaew (2025) warn that successful implementation requires structured
pedagogical frameworks to prevent overreliance on automated responses. Collectively, these studies
illustrate AI’s growing capacity to promote speaking development in EFL contexts, yet they primarily
focus on linguistic gains rather than learners’ affective responses—particularly how AI interactions
influence speaking anxiety and confidence in authentic communication.
Speaking anxiety
Speaking anxiety remains a major obstacle in EFL classrooms, often inhibiting learners from expressing
themselves confidently. Salsabil et al. (2025) found that AI voice chat interactions with ChatGPT
transformed silent learners into more confident speakers by providing a low-stress environment for oral
practice. This finding aligns with Celik et al. (2025), who observed that virtual speaking sessions
increased self- efficacy and reduced fear of negative evaluation. Likewise, Yildiz (2024) noted that
learners using ChatGPT reported less anxiety and greater engagement in oral communication.
Conversely, Poaquiza and EstefanÃa (2024) suggested that while AI voice chatbots lower initial anxiety,
the absence of human social cues can sometimes limit emotional connection and authentic interaction.
Nuñez et al. (2025) similarly indicated that although students became more fluent, some still doubted
the authenticity of AI feedback. Overall, these studies converge on the notion that AI voice tools have
the potential to mitigate speaking anxiety; however, the emotional and psychological dimensions of
communication, indicating a need for more context-specific exploration.
EFL and speaking development
Recent research confirms that AI-based applications, particularly ChatGPT, have emerged as promising
tools to enhance EFL learners’ speaking abilities. Celik et al. (2025) report that ChatGPT functions as
an effective virtual speaking tutor, significantly increasing learners’ self-efficacy and motivation to
speak. Similarly, Yildiz (2024) demonstrates that consistent ChatGPT use fosters greater linguistic
confidence and autonomy, especially when integrated into blended learning environments. Nuñez et al.
(2025) also observed measurable improvement in learners’ pronunciation, fluency, and coherence when
using ChatGPT’s voice function, emphasizing its potential for individualized oral feedback.
Complementing these findings, Huang (2024) shows that AI voice prompts can provide detailed, real-
time speaking feedback that teachers often cannot deliver due to time constraints. Nonetheless,
Kanoksilapatham and Takrudkaew (2025) warn that successful implementation requires structured
pedagogical frameworks to prevent overreliance on automated responses. Collectively, these studies
illustrate AI’s growing capacity to promote speaking development in EFL contexts, yet they primarily
focus on linguistic gains rather than learners’ affective responses—particularly how AI interactions
influence speaking anxiety and confidence in authentic communication.
Speaking anxiety
Speaking anxiety remains a major obstacle in EFL classrooms, often inhibiting learners from expressing
themselves confidently. Salsabil et al. (2025) found that AI voice chat interactions with ChatGPT
transformed silent learners into more confident speakers by providing a low-stress environment for oral
practice. This finding aligns with Celik et al. (2025), who observed that virtual speaking sessions
increased self- efficacy and reduced fear of negative evaluation. Likewise, Yildiz (2024) noted that
learners using ChatGPT reported less anxiety and greater engagement in oral communication.
Conversely, Poaquiza and EstefanÃa (2024) suggested that while AI voice chatbots lower initial anxiety,
the absence of human social cues can sometimes limit emotional connection and authentic interaction.
Nuñez et al. (2025) similarly indicated that although students became more fluent, some still doubted
the authenticity of AI feedback. Overall, these studies converge on the notion that AI voice tools have
the potential to mitigate speaking anxiety; however, the emotional and psychological dimensions of
pág. 4492
learners’ interaction with AI remain underexplored. This gap underscores the need for qualitative
inquiry into how EFL learners perceive and emotionally experience AI-assisted speaking practice.
AI voice assistant and accessibility for learners with visual impairment
AI-driven voice technology has become increasingly significant in supporting learners with visual
impairments by improving accessibility, autonomy, and engagement with educational content. Assistive
systems that integrate natural language processing, voice assistants, and speech-to-text/text-to-speech
features enable real-time interaction with digital environments that traditional screen readers alone
cannot provide (Al-Eidarous et al., 2024; SciVerse, 2025; Sustainable AI Solutions, 2025). Studies show
that Pa, thereby enhancing learning participation and reducing reliance on human assistance (SciVerse,
2025; Sustainable AI Solutions, 2025). These technologies facilitate autonomous learning by converting
complex text or environmental cues into accessible audio formats, which not only increases access to
information but also supports learners’ confidence and independence (SciVerse, 2025). Moreover,
emerging research underscores that AI-based assistive tools can provide personalized interaction via
adaptive speech modulation and real-time feedback, which may improve academic engagement and
participation among visually impaired students (SciVerse, 2025). However, challenges such as usability
issues, contextual understanding limitations, and ethical considerations remain, pointing to ongoing
development needs in inclusive AI design practices (Karamolegkou et al., 2025; SciVerse, 2025). These
insights highlight the transformative yet complex role of AI voice technology in accessible education
for learners with visual disabilities.
This study adopts a qualitative case study design to explore how AI voice interaction supports speaking
development and confidence in an adult EFL learner diagnosed with Stargardt disease. By focusing on
a single information-rich case, this research seeks to provide an in-depth understanding of how AI voice
technology functions as a mediational and assistive tool in real-life language learning practices. To
address the purpose of this study, the following research questions were formulated:
RQ1: How does the Gemini AI voice assistant support the speaking development of an adult EFL learner
with Stargardt disease?
RQ2: How does the learner perceive the role of AI voice interaction in enhancing her speaking
confidence and autonomy?
learners’ interaction with AI remain underexplored. This gap underscores the need for qualitative
inquiry into how EFL learners perceive and emotionally experience AI-assisted speaking practice.
AI voice assistant and accessibility for learners with visual impairment
AI-driven voice technology has become increasingly significant in supporting learners with visual
impairments by improving accessibility, autonomy, and engagement with educational content. Assistive
systems that integrate natural language processing, voice assistants, and speech-to-text/text-to-speech
features enable real-time interaction with digital environments that traditional screen readers alone
cannot provide (Al-Eidarous et al., 2024; SciVerse, 2025; Sustainable AI Solutions, 2025). Studies show
that Pa, thereby enhancing learning participation and reducing reliance on human assistance (SciVerse,
2025; Sustainable AI Solutions, 2025). These technologies facilitate autonomous learning by converting
complex text or environmental cues into accessible audio formats, which not only increases access to
information but also supports learners’ confidence and independence (SciVerse, 2025). Moreover,
emerging research underscores that AI-based assistive tools can provide personalized interaction via
adaptive speech modulation and real-time feedback, which may improve academic engagement and
participation among visually impaired students (SciVerse, 2025). However, challenges such as usability
issues, contextual understanding limitations, and ethical considerations remain, pointing to ongoing
development needs in inclusive AI design practices (Karamolegkou et al., 2025; SciVerse, 2025). These
insights highlight the transformative yet complex role of AI voice technology in accessible education
for learners with visual disabilities.
This study adopts a qualitative case study design to explore how AI voice interaction supports speaking
development and confidence in an adult EFL learner diagnosed with Stargardt disease. By focusing on
a single information-rich case, this research seeks to provide an in-depth understanding of how AI voice
technology functions as a mediational and assistive tool in real-life language learning practices. To
address the purpose of this study, the following research questions were formulated:
RQ1: How does the Gemini AI voice assistant support the speaking development of an adult EFL learner
with Stargardt disease?
RQ2: How does the learner perceive the role of AI voice interaction in enhancing her speaking
confidence and autonomy?
pág. 4493
METHODS
This study adopts a qualitative instrumental case study design to explore how AI voice interaction
functions as an assistive tool for enhancing EFL speaking skills and confidence. A case study approach
is appropriate because it allows for an in depth examination of a contemporary phenomenon within its
real life context. Rather than seeking generalizable findings, this design aims to generate a rich,
contextualized understanding of one information rich case.
The study focuses on a single adult learner diagnosed with Stargardt disease, a degenerative retinal
condition that affects central vision. By examining her experiences with AI voice interaction, the study
seeks to understand how voice-based technologies mediate speaking development, autonomy, and
confidence in the context of visual impairment.
Participant
The participant is a 34-year-old Ecuadorian adult originally from Manta, Ecuador, who is currently
residing in Manhattan, Kansas, United States. She was diagnosed with Stargardt disease at the age of 20
and currently retains approximately 50% of her vision. Due to the progressive nature of the condition,
she relies increasingly on auditory-based technologies in her daily life.
She is an English as a Foreign Language (EFL) learner who actively uses AI voice assistants, particularly
Gemini, to practice and improve her English-speaking skills. The participant was selected through
purposive sampling as an information-rich case because of her sustained engagement with AI voice
interaction and her lived experience navigating language learning with a visual impairment.
To protect confidentiality, a pseudonym will be used in reporting the findings, and identifying personal
details will be omitted.
Instrument and materials
The primary data collection instrument was a semi-structured interview protocol designed to explore the
participant’s experiences using AI voice interaction for speaking development. The semi-structured
format allowed flexibility to probe deeper into emerging themes while maintaining alignment with the
research questions.
The interview included open-ended questions related to: (a) frequency and contexts of AI voice use, (b)
perceived impact on fluency and pronunciation, (c) emotional responses and speaking confidence, (d)
METHODS
This study adopts a qualitative instrumental case study design to explore how AI voice interaction
functions as an assistive tool for enhancing EFL speaking skills and confidence. A case study approach
is appropriate because it allows for an in depth examination of a contemporary phenomenon within its
real life context. Rather than seeking generalizable findings, this design aims to generate a rich,
contextualized understanding of one information rich case.
The study focuses on a single adult learner diagnosed with Stargardt disease, a degenerative retinal
condition that affects central vision. By examining her experiences with AI voice interaction, the study
seeks to understand how voice-based technologies mediate speaking development, autonomy, and
confidence in the context of visual impairment.
Participant
The participant is a 34-year-old Ecuadorian adult originally from Manta, Ecuador, who is currently
residing in Manhattan, Kansas, United States. She was diagnosed with Stargardt disease at the age of 20
and currently retains approximately 50% of her vision. Due to the progressive nature of the condition,
she relies increasingly on auditory-based technologies in her daily life.
She is an English as a Foreign Language (EFL) learner who actively uses AI voice assistants, particularly
Gemini, to practice and improve her English-speaking skills. The participant was selected through
purposive sampling as an information-rich case because of her sustained engagement with AI voice
interaction and her lived experience navigating language learning with a visual impairment.
To protect confidentiality, a pseudonym will be used in reporting the findings, and identifying personal
details will be omitted.
Instrument and materials
The primary data collection instrument was a semi-structured interview protocol designed to explore the
participant’s experiences using AI voice interaction for speaking development. The semi-structured
format allowed flexibility to probe deeper into emerging themes while maintaining alignment with the
research questions.
The interview included open-ended questions related to: (a) frequency and contexts of AI voice use, (b)
perceived impact on fluency and pronunciation, (c) emotional responses and speaking confidence, (d)
pág. 4494
the role of voice interaction in accessibility and autonomy.
The interview lasted approximately 20 minutes and was conducted in Spanish to ensure comfort and
depth of expression. With informed consent, the interview was recorded and later transcribed for
thematic analysis.
Data collection procedures
The participant was contacted through academic referral and voluntarily agreed to participate in the
study. The interview was conducted individually via WhatsApp due to geographical distance (the
participant resides in Manhattan, Kansas). The session lasted approximately 20 minutes and was
recorded with permission.
Following the interview, the recording was transcribed verbatim. The transcript was reviewed multiple
times to ensure accuracy and to facilitate thematic coding.
Ethical considerations
Ethical approval procedures were followed in accordance with institutional research guidelines. The
participant was fully informed of the voluntary nature of the study and her right to withdraw at any time
without consequences.
Confidentiality was ensured through the use of a pseudonym and the removal of identifying details.
Given the inclusion of a medical condition (Stargardt disease), special care was taken to present
information respectfully and in a non-stigmatizing manner.
All digital files, including audio recordings and transcripts, are stored on a password-protected device
accessible only to the researcher. The study poses minimal risk; however, the participant was informed
that she could decline to answer any question that caused discomfort.
FINDINGS
The analysis of the interview transcript revealed five central themes regarding the use of an AI voice
assistant (Gemini) for English learning. These themes reflect (1) accessibility, (2) AI-mediated informal
speaking practice, (3) autonomy and metacognitive regulation, (4) reduced anxiety in AI interaction,
and (5) perceived limitations of AI mediated learning. Table 1 summarizes the thematic interpretation
derived from the participant’s responses.
Accessibility as linguistic empowerment
the role of voice interaction in accessibility and autonomy.
The interview lasted approximately 20 minutes and was conducted in Spanish to ensure comfort and
depth of expression. With informed consent, the interview was recorded and later transcribed for
thematic analysis.
Data collection procedures
The participant was contacted through academic referral and voluntarily agreed to participate in the
study. The interview was conducted individually via WhatsApp due to geographical distance (the
participant resides in Manhattan, Kansas). The session lasted approximately 20 minutes and was
recorded with permission.
Following the interview, the recording was transcribed verbatim. The transcript was reviewed multiple
times to ensure accuracy and to facilitate thematic coding.
Ethical considerations
Ethical approval procedures were followed in accordance with institutional research guidelines. The
participant was fully informed of the voluntary nature of the study and her right to withdraw at any time
without consequences.
Confidentiality was ensured through the use of a pseudonym and the removal of identifying details.
Given the inclusion of a medical condition (Stargardt disease), special care was taken to present
information respectfully and in a non-stigmatizing manner.
All digital files, including audio recordings and transcripts, are stored on a password-protected device
accessible only to the researcher. The study poses minimal risk; however, the participant was informed
that she could decline to answer any question that caused discomfort.
FINDINGS
The analysis of the interview transcript revealed five central themes regarding the use of an AI voice
assistant (Gemini) for English learning. These themes reflect (1) accessibility, (2) AI-mediated informal
speaking practice, (3) autonomy and metacognitive regulation, (4) reduced anxiety in AI interaction,
and (5) perceived limitations of AI mediated learning. Table 1 summarizes the thematic interpretation
derived from the participant’s responses.
Accessibility as linguistic empowerment