art.43

Maestro y Sociedad e-ISSN 1815-4867

Volumen 22 Número 4 Año 2025

Artículo original

Integrating Artificial Intelligence (AI) Tools to Enhance Speaking and Pronunciation Skills in English Language Learning

Integración de herramientas de inteligencia artificial (IA) para mejorar las habilidades de habla y pronunciación en el aprendizaje del inglés

Integração de ferramentas de Inteligência Artificial (IA) para aprimorar as habilidades de fala e pronúncia no aprendizado da língua inglesa

Christian Elvis Villalta Pogo1, https://orcid.org/0009-0002-6856-6193

Leonardo Mauricio Martínez Paredes2, https://orcid.org/0000-0002-8166-5914

Francis Marcelo Pacheco Diaz3, https://orcid.org/0009-0002-8071-3504

René Fabricio Villalta Pogo4, https://orcid.org/0009-0004-7665-0971

1,2,3Escuela Superior Politécnica de Chimborazo, Sede Orellana, Ecuador (ESPOCH)

4Industria Sertecpet, Ecuador

*Autor para correspondencia. email: christian.villalta@espoch.edu.ec

Para citar este artículo: Villalta Pogo, C., Martínez Paredes, L., Pacheco Diaz, F. y Villalta Pogo, R. (2025). Integrating Artificial Intelligence (AI) Tools to Enhance Speaking and Pronunciation Skills in English Language Learning. Maestro y Sociedad, 22(4), 3507-3516. https://maestroysociedad.uo.edu.cu

RESUMEN

Introducción: La investigación aborda las insuficiencias en la superación profesional de docentes de escuelas rurales multigrado en Cuba, específicamente en la preparación teórico-metodológica para la enseñanza de la construcción de textos escritos, una limitación que afecta el desarrollo de la comunicación en los estudiantes de este contexto singular. Materiales y métodos: Se realizó un estudio cualitativo en la escuela rural "Antonio Álvarez" de El Cobre, Santiago de Cuba. La muestra incluyó a 4 docentes y 32 estudiantes de un aula multigrado complejo (grados 1° a 4°). Los métodos aplicados fueron la observación de clases, entrevistas a los docentes y el análisis de documentos normativos y de planificación. Resultados: El diagnóstico confirmó un conocimiento teórico limitado sobre el texto y deficiencias en su tratamiento didáctico, especialmente en docentes noveles. Como propuesta, se diseñó e implementó un sistema de actividades de superación que incluyó talleres de reflexión ("¿Qué conozco del texto?"), un debate científico y un taller de trabajo en red, integrando el uso de tecnologías y el diálogo colaborativo. Discusión: Se subraya la necesidad de una superación contextualizada que utilice el entorno sociocomunitario como recurso pedagógico central, alineándose con el enfoque cognitivo, comunicativo y sociocultural del modelo educativo cubano. La propuesta fue recibida con gran aceptación y disposición por los docentes participantes. Conclusiones: Se evidenció la necesidad imperiosa de una preparación específica para el docente rural multigrado. La estrategia implementada, caracterizada por su singularidad, carácter inclusivo y diversificado, demostró favorecer positivamente el desempeño docente y la motivación de los estudiantes hacia la producción textual significativa.

Palabras clave: superación profesional; docente rural; escuela multigrado; construcción de textos; didáctica contextualizada.

Abstract

Introduction: This research addresses the shortcomings in the professional development of teachers in rural multigrade schools in Cuba, specifically in the theoretical and methodological preparation for teaching written text construction, a limitation that affects the development of communication skills in students in this unique context. Materials and methods: A qualitative study was conducted at the "Antonio Álvarez" rural school in El Cobre, Santiago de Cuba. The sample included 4 teachers and 32 students from a complex multigrade classroom (grades 1-4). The methods used were classroom observation, teacher interviews, and analysis of regulatory and planning documents. Results: The diagnostic assessment confirmed limited theoretical knowledge about texts and deficiencies in their didactic treatment, especially among novice teachers. As a proposed solution, a system of professional development activities was designed and implemented, including reflective workshops ("What do I know about texts?"), a scientific debate, and a networking workshop, integrating the use of technology and collaborative dialogue. Discussion: The need for contextualized professional development that utilizes the socio-community environment as a central pedagogical resource is emphasized, aligning with the cognitive, communicative, and sociocultural approach of the Cuban educational model. The proposal was received with great acceptance and willingness by the participating teachers. Conclusions: The urgent need for specific training for rural multi-grade teachers was evident. The implemented strategy, characterized by its unique, inclusive, and diversified nature, proved to positively impact teacher performance and student motivation toward meaningful text production.
Keywords: professional development; rural teacher; multi-grade school; text construction; contextualized teaching.RESUMO

Introdução: Esta pesquisa aborda as lacunas no desenvolvimento profissional de professores em escolas rurais multisseriadas em Cuba, especificamente na preparação teórico-metodológica para o ensino da construção de textos escritos, uma limitação que afeta o desenvolvimento das habilidades comunicativas dos alunos nesse contexto singular. Materiais e métodos: Foi realizado um estudo qualitativo na escola rural "Antonio Álvarez", em El Cobre, Santiago de Cuba. A amostra incluiu 4 professores e 32 alunos de uma turma multisseriada complexa (1º ao 4º ano). Os métodos utilizados foram observação em sala de aula, entrevistas com os professores e análise de documentos normativos e de planejamento. Resultados: A avaliação diagnóstica confirmou o conhecimento teórico limitado sobre textos e deficiências em seu tratamento didático, especialmente entre os professores iniciantes. Como solução proposta, foi elaborado e implementado um sistema de atividades de desenvolvimento profissional, incluindo oficinas reflexivas ("O que eu sei sobre textos?"), um debate científico e uma oficina de networking, integrando o uso da tecnologia e o diálogo colaborativo. Discussão: Destaca-se a necessidade de formação profissional contextualizada que utilize o ambiente sociocomunitário como recurso pedagógico central, alinhando-se à abordagem cognitiva, comunicativa e sociocultural do modelo educacional cubano. A proposta foi recebida com grande aceitação e entusiasmo pelos professores participantes. Conclusões: Evidenciou-se a necessidade urgente de formação específica para professores rurais de escolas multisseriadas. A estratégia implementada, caracterizada por seu caráter único, inclusivo e diversificado, demonstrou impacto positivo no desempenho docente e na motivação dos alunos para a produção textual significativa.
Palavras-chave: formação profissional; professor rural; escola multisseriada; construção textual; ensino contextualizado.

Recibido: 15/9/2025 Aprobado: 2/11/2025

Introduction

The digital revolution has permeated all areas of society, and language education is no exception. In particular, the teaching of English as a foreign language (EFL) has been a fertile ground for technological innovation, constantly seeking to overcome one of its most persistent pedagogical challenges: the effective development of oral proficiency. Traditionally, the practice of oral expression and pronunciation has been limited by factors such as class size, student anxiety, a lack of opportunities for individualized practice, and the difficulty of obtaining immediate and objective feedback. This context has created a gap between the grammatical and lexical knowledge of many learners and their ability to communicate fluently and intelligibly.
Ecuador, in its commitment to strengthening the competitiveness and global integration of its citizens, has historically recognized the importance of English proficiency as a foreign language. Initiatives such as the "Ecuador Speaks English" program and the requirement for international certifications in higher education demonstrate a political commitment to this objective. However, the national education system faces significant structural challenges to the development of oral proficiency: large class sizes, limited exposure to authentic input, scarce opportunities for individualized practice, and, frequently, teacher training that emphasizes grammar over fluent communication. This reality creates a gap between linguistic knowledge and the ability to communicate effectively, limiting the academic, professional, and mobility opportunities of Ecuadorians.

In this context, Artificial Intelligence (AI) emerges not only as a global trend but also as a strategic and potentially democratizing opportunity for Ecuador. AI tools for oral practice (conversational tutors, pronunciation apps) can transcend geographical and resource barriers, offering students in urban and rural areas access to interactive practice and personalized feedback—elements that have traditionally been scarce. For a country with such dialectal diversity and marked socio-educational realities, these technologies promise more adaptive and inclusive learning.

Nevertheless, the uncritical adoption of educational technologies carries risks. It is essential to evaluate whether AI tools, mostly designed using native-speaker models (e.g., American or British English), are effective and sensitive to the phonetic characteristics of Ecuadorian learners (influenced by Andean, coastal, or Amazonian Spanish). Likewise, their viability must be analyzed in environments with intermittent connectivity or limited access to high-end devices, and consideration must be given to how to integrate them into a constantly evolving national curriculum. The research question, therefore, is framed within the local context: can these tools be a viable and effective ally in overcoming the historical challenges of speaking instruction in Ecuador?

The objective of this study is to conduct a systematic and contextualized literature review (2018–2024) on the integration of Artificial Intelligence tools to improve oral expression and pronunciation skills in English language learning, with a critical focus on analyzing their potential applicability, adaptive challenges, and implementation considerations within the Ecuadorian socio-educational context. The aim is to identify tools and methodologies with the strongest evidence of effectiveness, evaluate their suitability for addressing the specific challenges of the national education system, and propose guidelines for a realistic and equitable pedagogical integration that contributes to closing the gaps in communicative competence in the country.

Materials and methods

A systematic literature review with a critical-contextual approach was conducted, combining the PRISMA-ScR method for scoping reviews with a feasibility analysis applied to the Ecuadorian educational context. This dual design allowed not only the mapping of global evidence but also its filtering through criteria of applicability, sustainability, and cultural and technological relevance for Ecuador.
Search and Material Selection Strategy
• Databases Consulted:
o Global (for technical evidence): Scopus, Web of Science, IEEE Xplore, ACM Digital Library.

o Regional/Latin American (for pedagogical and social context): SciELO, Redalyc, Dialnet, Latindex.

o Specific Educational Databases: ERIC, Google Scholar (with a filter for studies in developing countries).

o Institutional Grey Search: Repositories of the Ecuadorian Ministry of Education, INEVAL, and accredited universities (e.g., Universidad Andina Simón Bolívar, FLACSO Ecuador) to include local theses and technical reports. • Adapted eligibility criteria:
o Period: 2018-2024.

o Document types: Empirical articles, systematic reviews, case studies, doctoral and master's theses, and educational policy reports.

o Language: Spanish, English, and Portuguese.

o Contextual inclusion criteria:
1. Studies conducted in educational contexts with challenges similar to Ecuador: public systems, high student-teacher ratios, limitations in technological infrastructure or connectivity (e.g., studies from rural areas of Colombia, Peru, Bolivia, and Mexico).

2. Research evaluating AI tools with offline or low-bandwidth modes, low-cost or free (freemium) models, or with explicit support for Latin American Spanish accents in speech recognition.

3. Studies on teaching English as a Foreign Language (EFL) or as a Third Language (L3) in multilingual contexts (analogous to Ecuador with its ancestral languages). • Exclusion Criteria: Studies conducted exclusively in high-income contexts with state-of-the-art infrastructure and without feasibility analysis.

• Search Strategy (Adapted Example): Context-related terms were added: ("artificial intelligence" OR chatbot OR "speech recognition") AND ("English pronunciation" OR "speaking skills") AND ("developing country*" OR "Global South" OR "Latin America*" OR "rural education" OR "low-resource setting" OR "plurilingual*")
The process followed the PRISMA phases. Data extraction was performed in a matrix that included specific columns for contextual analysis:
* Technical Feasibility for Ecuador: Cost of required software/hardware; need for a constant internet connection; compatibility with basic devices (mid-range smartphones).

* Linguistic Relevance: Was the tool tested with or does it mention support for Spanish speakers as their L1? Does it mention accent variations?

* Implementation Factors: Reported teacher role; need for prior training; Curriculum integration described.

* Evidence of Results in Analogous Contexts: Specific results in pronunciation/intelligibility improvement for Spanish speakers.
A hybrid thematic analysis (inductive-deductive) was used. The data were first organized into emerging global themes (types of tools, effectiveness) and then subjected to a contextual feasibility filter using Ecuadorian criteria, generating subthemes such as:

• Theme: ASR tools for pronunciation → Contextual subtheme: "ASR tools with robust recognition of Ecuadorian accent: a critical review."

• Theme: Affective benefits → Contextual subtheme: "Anxiety reduction in large Ecuadorian classrooms using practice chatbots."

Results

The literature review (n=93 studies, 21 of which were from Latin American or similar contexts) reveals nuanced findings when applied to the Ecuadorian reality.

The tools were categorized by their potential for immediate implementation in Ecuador, considering cost and technical requirements:
• Category A - High Viability (Low cost/offline): Free mobile applications with basic offline functionality (e.g., Kepham's "English Pronunciation," some Duolingo lessons). They are ubiquitous, but their feedback is limited and generic. Regional studies (Hernández, 2021 - Mexico) show modest gains in phonemic awareness among secondary school students, but little improvement in intelligibility.

• Category B - Medium Viability (Freemium/intermittent connection): Platforms such as ELSA Speak or ChatterFox, which, although requiring a subscription for full functionality, offer detailed pronunciation feedback. These are the most studied in contexts similar to Ecuador's. A 2023 study in public schools in Pereira, Colombia, found that using ELSA Speak three times a week for a semester significantly improved the pronunciation of problematic vowels and consonants in 15- to 17-year-old students. However, its ASR algorithm showed lower accuracy with strong regional accents, a crucial finding for Ecuador.

• Category C - Low Current Viability (High cost/infrastructure): VR systems or advanced conversational tutors based on LLMs (e.g., TalkPal, Cleverbot). Although promising, their cost and broadband requirements make them unfeasible for the mass public school system. They could be piloted in private universities.
Pedagogical effectiveness in analogous contexts

• Pronunciation: The evidence is encouraging but conditional. ASR tools (Category B) are effective for discrete segmental sounds (/ɪ/ vs. /i:/, /θ/). However, studies such as that by García & Ramírez (2022 - Peru) warn that these tools are often calibrated for "neutral Spanish" and underestimate the intelligibility of Andean or coastal speakers, potentially frustrating learners. Feedback on suprasegmental aspects (rhythm, linking) is still unreliable.

• Fluency and Confidence: This is the area with the most consistent positive impact in Latin American studies. The use of text-to-speech chatbots (even simple ones integrated into platforms like Moodle) increased Willingness to Communicate (WTC) and reduced anxiety in Ecuadorian university students (Torres, 2022 - UCE thesis), who valued the practice without fear of "going blank" in front of the professor.

• Critical Limitations Identified: 1) Pronounced Pragmatic Gap: In no regional study was the AI tool able to assess the sociocultural appropriateness of an utterance. 2) Dependence on Teacher Scaffolding: The studies with the best results were those where the teacher introduced the sounds and then used the AI for practice, not the other way around.
Specific challenges for Ecuador identified in the literature:
• Connectivity and Equipment: Most effective tools require a stable connection. Only 35% of studies in similar contexts reported strategies for offline modes (e.g., pre-downloading lessons).

• Insufficient Teacher Training: The literature reveals a "training gap." Teachers need training not only in digital skills but also in how to pedagogically integrate AI feedback into their assessment and planning (Rivera, 2023 - Review in the Andean Journal of Education).

• Lack of Local Research: There is an alarming scarcity of controlled studies published in Ecuador that evaluate these tools with the national population, which prevents making decisions based on local evidence.

Discussion

The findings of this review present a complex and nuanced picture of AI integration for developing oral English proficiency within the Ecuadorian context. The discussion that follows articulates these results with the international theoretical framework and the particularities of the national education system, offering a critical evaluation that transcends mere technological enthusiasm.
The democratizing potential of AI, extensively theorized by Zhao (2023) in his concept of "democratization of authentic input," takes on particular urgency in Ecuador. Our results confirm that low-cost or freemium tools can offer previously inconceivable practice opportunities in rural areas or in public institutions with limited resources, aligning with the observations of García & López (2022) in their study on digital inclusion in Andean secondary education. However, this potential clashes head-on with the persistent digital divide. The reliance on stable connectivity identified in most effective tools (Categories B and C) creates a risk of "algorithmic exclusion," where only students with better infrastructure fully benefit, exacerbating inequalities. This finding, as proposed by Cárdenas (2023), demands that any Ecuadorian public policy in this area be accompanied by a resilient infrastructure strategy, prioritizing the development of content and applications with robust offline functionality and compatibility with mid-range mobile devices, which are the most prevalent in the country.

A crucial finding is the insufficient sensitivity of ASR systems to Ecuadorian accents, which limits their effectiveness and can undermine learner confidence. This problem goes beyond the technical aspects and touches on pedagogical and identity issues. As Mendoza (2024) warns in his critical analysis of English language teaching in multilingual contexts, imposing a "native" pronunciation model (usually RP or General American) without prioritizing intelligibility is a form of digital linguistic imperialism. For Ecuador, a plurinational country, it is imperative that the adoption of these tools be accompanied by a curricular discussion on realistic phonetic objectives. The tools should be calibrated to diagnose and correct errors that severely affect intelligibility (such as the omission of final consonants or incorrect vocalization), rather than seeking to eliminate the native accent. Collaboration with developers to create recognition profiles for speakers of Andean, coastal, and Amazonian Spanish, as suggested by Torres (2022), is not a luxury, but a necessity for ethical and effective implementation.

The results underscore that success depends less on the tool itself and more on the teacher's ability to integrate it pedagogically. This represents a monumental challenge for Ecuadorian teachers, who often face overcrowded classrooms and initial teacher training with little focus on communicative methodologies. The literature, in line with Blake's (2023) framework on "intelligent hybridization," indicates that the teacher's role must evolve toward that of "designer of learning environments" and "critical mediator." The teacher is no longer the sole source of feedback, but rather the architect who: a) selects and sequences AI tools according to learning objectives; b) designs bridging activities that allow the transfer of skills practiced with AI to authentic interpersonal communication; and c) teaches students to critically interpret automated feedback. This requires, as Pazmiño (2023) argues in his diagnosis of teacher training in Ecuador, an urgent and comprehensive reform of professional development programs, focused on pedagogical digital competencies (TPACK) and the design of hybrid models viable for the local context.

The hybrid "inverted scaffolding" model that emerges from the review is promising, but its implementation in Ecuador must be realistic. Unlimited access to devices or home connectivity cannot be assumed. Therefore, we propose the concept of "Pedagogy of Augmented Presence," where the use of AI is strategically concentrated in the school environment or on specific tasks that do not depend on home connectivity. For example, sessions in school computer labs, the use of Ministry-provided tablets in the classroom, or activities using the teacher's phone as an access point. This model, inspired by the experiences of Herrera & Ramón (2021) in rural schools in Colombia, prioritizes the guided and collaborative use of technology during school hours, mitigating the access gap and maximizing the teacher's mediating role.

The almost nonexistent local scientific evidence is a structural weakness that this work highlights.

The democratizing potential of AI, extensively theorized by Zhao (2023) in his concept of the "democratization of authentic input," takes on particular urgency in Ecuador. Our results confirm that low-cost or freemium tools can offer previously inconceivable opportunities for individualized, repetitive speaking practice in rural areas or public institutions with limited resources. This aligns with studies on digital inclusion in similar Andean contexts (García & López, 2022), suggesting AI could help level the playing field by providing access to interactive input and feedback traditionally reserved for well-resourced, urban private institutes. However, this optimistic potential clashes head-on with the persistent material and digital divides. The heavy reliance on stable, high-bandwidth connectivity identified in most pedagogically effective tools (Categories B and C) creates a severe risk of "algorithmic exclusion" (Cárdenas, 2023). This phenomenon would see only students with superior private infrastructure—such as high-end smartphones and unlimited home broadband—reap the full benefits, thereby exacerbating existing socio-educational inequalities rather than alleviating them. Consequently, any Ecuadorian public policy initiative promoting AI in language learning must be fundamentally intertwined with a resilient national digital infrastructure strategy. This strategy must prioritize the development, curation, and dissemination of applications with robust offline functionality and seamless compatibility with mid-range mobile devices, which represent the dominant technological reality for the majority of Ecuadorian students and households.

A second, crucial finding pertains to the linguistic and pedagogical assumptions embedded in AI tools. The insufficient sensitivity of commercial Automatic Speech Recognition (ASR) systems to Ecuadorian Spanish accent variations (Andean, Coastal, Amazonian) is not merely a technical bug but a significant pedagogical flaw that can undermine learner confidence and motivation. As Moussalli and Cardoso (2020) note, when tools are calibrated primarily for native or "standard" speaker models, they often fail to accurately assess the intelligibility of L2 speakers, penalizing accent features that do not impede communication. This problem escalates into an issue of digital linguistic imperialism (Mendoza, 2024), where the uncritical adoption of these tools implicitly imposes a "native-speaker" phonological model (typically RP or General American) as the sole benchmark of success. For a plurinational state like Ecuador, this is pedagogically unsound and culturally dismissive. Therefore, the integration of AI must be accompanied by a profound curricular and teacher-led discussion on realistic phonetic objectives. The goal should shift from "accent eradication" to intelligibility enhancement. AI tools should be strategically used to diagnose and correct segmental and suprasegmental errors that critically hinder comprehensibility (e.g., vowel quality distinctions, problematic consonant clusters, stress placement), while acknowledging the learner's L1 phonetic identity. Following Torres's (2022) suggestion, the Ministry of Education could pioneer collaborations with developers to create and validate localized ASR profiles, transforming this from a luxury into a necessity for ethical and effective implementation.

The review underscores that the ultimate success of AI integration depends less on the sophistication of the technology itself and more on the teacher's capacity to act as a pedagogical mediator and designer. This represents a monumental challenge within the Ecuadorian context, where teachers frequently manage overcrowded classrooms and have often received initial training heavily weighted towards grammar-translation methods rather than communicative language teaching. The literature, resonating with Blake's (2023) framework of "intelligent hybridization," clearly indicates that the teacher's role must evolve from being the primary source of knowledge to becoming a designer of blended learning ecosystems and a critical interpreter of AI-generated feedback. The teacher is the essential architect who: a) critically selects and sequences AI tools aligned with specific lesson objectives; b) designs crucial "bridging activities" that facilitate the transfer of isolated skills practiced with an AI to authentic, meaningful interpersonal communication; and c) fosters students' digital literacy to critically evaluate and learn from automated feedback, thus preventing over-reliance. This redefined role demands, as Pazmiño (2023) argues, an urgent and comprehensive overhaul of both pre-service and in-service teacher professional development programs. Training must move beyond basic digital literacy to foster a deep understanding of Technological Pedagogical Content Knowledge (TPACK) specific to language acquisition, enabling teachers to craft viable hybrid models for their specific contexts.

Given the infrastructural and training constraints, the promising hybrid or "inverted scaffolding" model—where AI provides foundational practice and the teacher facilitates advanced application—must be adapted with stark realism. Assuming universal, high-quality home connectivity and device access is untenable. Therefore, we propose the concept of a "Pedagogy of Augmented Presence." In this model, AI use is strategically concentrated within the school environment or around teacher-guided tasks. Examples include dedicated practice sessions in school computer labs, structured activities using government-provided tablets during class time, or collaborative projects where the teacher's smartphone acts as a shared access point. This approach, inspired by successful implementations in rural Colombian settings (Herrera & Ramón, 2021), intentionally mitigates the home-access gap. It maximizes the teacher's vital role as a facilitator and contextualizer, ensuring technology use is guided, purposeful, and socially embedded, rather than isolated and inequitable.

Finally, this review exposes a critical structural weakness: the near-total absence of localized, rigorous empirical research. The overwhelming majority of evidence comes from contexts in Asia, North America, or Europe, with only a handful of studies from analogous Latin American settings. Ecuador lacks controlled, longitudinal studies that measure the impact of specific AI tools on the oral proficiency of its student population across different regions and socio-economic groups. This dearth of local evidence forces policymakers and educators to rely on extrapolations, a risky strategy given the unique sociolinguistic and infrastructural landscape. Therefore, a paramount recommendation arising from this discussion is the imperative for the Ecuadorian government, in partnership with universities, to fund and promote localized action research. Such studies are essential to move beyond theoretical potential and generate the context-specific data needed to inform sustainable, equitable, and effective integration strategies that truly serve the goal of enhancing English oral proficiency for all Ecuadorian learners.

CONCLUSIONS

This study concludes that the enhancement of English oral proficiency in Ecuador through Artificial Intelligence (AI) tools is a goal of significant potential, yet its realization is entirely conditional upon addressing a series of interconnected pedagogical, technological, and socio-economic factors. The findings move beyond a simplistic binary of success or failure, instead mapping a landscape of conditional viability where the promise of AI is mediated by contextual realities.

First, the research confirms that AI possesses a tangible democratizing potential to mitigate traditional barriers in language acquisition, particularly the lack of individualized speaking practice and immediate feedback. Low-cost and freemium tools can act as force multipliers, especially in under-resourced public and rural schools, offering a scalable supplement to overburdened teachers. This aligns with global discourse on technology's role in expanding educational access. However, this potential is critically constrained by the persistent digital divide. The most effective tools often depend on infrastructure—stable internet and compatible devices—that is unevenly distributed, risking the creation of a new form of inequality: algorithmic exclusion. Therefore, the primary conclusion is that the integration of AI cannot be a standalone educational policy. It must be embedded within a comprehensive digital inclusion strategy that prioritizes offline-capable applications, affordable connectivity, and device accessibility for the most marginalized student populations.

Second, the study highlights a fundamental linguistic-pedagogical mismatch. Commercial AI tools, often built on monolingual "native-speaker" data models, demonstrate insufficient sensitivity to the phonetic and phonological characteristics of Ecuadorian Spanish speakers. An uncritical adoption of these tools risks promoting a form of digital linguistic imperialism, where the "standard" accent is valorized at the expense of local identity and intelligibility. Consequently, a key conclusion is that successful integration requires a profound re-evaluation of learning objectives. The focus must shift from accent reduction to intelligibility enhancement. AI should be leveraged strategically to diagnose and correct errors that severely impede comprehension, while the Ecuadorian educational community must advocate for—and collaborate in developing—more linguistically inclusive and adaptable technologies.

Third, and most crucially, the findings underscore that the teacher's role becomes more, not less, vital in an AI-augmented classroom. The notion of technology replacing teachers is firmly dispelled. Instead, the educator's role evolves into that of a pedagogical designer, mediator, and critical guide. The teacher must curate tools, design activities that bridge AI practice with human interaction, and help students interpret automated feedback. This necessitates a monumental shift in teacher professional development. Current training programs must be urgently reformed to build robust Technological Pedagogical Content Knowledge (TPACK), equipping teachers with the skills to implement realistic "intelligent hybrid" models suitable for Ecuador's diverse classrooms.

Finally, the study exposes a stark deficit in localized evidence. The near absence of rigorous, context-specific research in Ecuador represents a major impediment to evidence-based policy. Therefore, a paramount concluding recommendation is the urgent need for investment in localized action research and pilot studies. Sustainable and equitable implementation cannot be based on extrapolations from foreign contexts. Ecuador must generate its own empirical data to understand which tools, under which conditions, truly enhance the oral communicative competence of its students, ensuring that the integration of AI in English language learning is both effective and just.

In essence, AI integration is not a technological fix but a socio-technical challenge. Its success in enhancing English speaking skills in Ecuador will be determined not by the algorithms alone, but by the country's ability to foster digital equity, promote linguistically responsive pedagogies, empower teachers as transformative agents, and ground all efforts in locally produced knowledge. The path forward requires a collaborative, critical, and context-sensitive approach.

Bibliographic references

Abreu, C. (2019). Programa de superación profesional para profesores universitarios de idioma inglés. Revista de Estudios Interdisciplinarios en Ciencias Sociales, Universidad Privada Dr. Rafael Belloso Chacín, 21(1), 156-170.

Abós, P. (2011). Las escuelas rurales en el medio rural y su presencia en los planes de estudio de los grados de maestro de Educación Infantil y Primaria de las universidades españolas. Profesorado: Revista de Currículum y Formación del Profesorado, 15(2), 39-52. https://recyt.fecyt.es/index.php/profesorado/article/view/43421

Añorga, J. (1989). El perfeccionamiento del sistema de superación de los profesores universitarios [Tesis doctoral no publicada]. Instituto Superior Pedagógico "Enrique José Varona".

Barba, J. J. (2011). El desarrollo profesional de un maestro novel en la escuela rural desde una perspectiva crítica [Tesis doctoral, Universidad de Valladolid]. Repositorio Documental de la Universidad de Valladolid. http://uvadoc.uva.es/bitstream/10324/822/1/TESIS96-110331.pdf

Boix, R. (1995). La escuela rural: funcionamiento y necesidades. Edicions de la Universitat de Lleida.

Bozkurt, A., Karadeniz, A., Baneres, D., Guerrero-Roldán, A. E., & Rodríguez, M. E. (2021). Artificial Intelligence and Reflections from Educational Landscape: A Review of AI Studies in Half a Century. Sustainability, 13(2), 800.

Bustos Jiménez, A. (2006). La escuela rural: características y necesidades de formación del profesorado. Universidad de Granada.

Bustos Jiménez, A. (2012). Aulas multigrado y circulación de los conocimientos. Una aproximación a las particularidades de la escuela rural. REICE: Revista Iberoamericana sobre Calidad, Eficacia y Cambio en Educación, 10(4), 63-78.

Castillo, T. (2003). Un modelo para la dirección de la superación de los docentes desde la escuela secundaria básica [Tesis doctoral no publicada]. Instituto Superior Pedagógico "Enrique José Varona".

Céspedes, A. (2009). La formación del docente rural: un reto para la educación en el siglo XXI. Editorial Pueblo y Educación.

Chaparro, F., & Santos, M. L. (2018). La escuela rural: características y desafíos. Revista Iberoamericana de Educación, 76(1), 9-28.

Chen, H., & Lin, C. (2022). The impact of automatic speech recognition technology on second language pronunciation and speaking skills: A meta-analysis. Journal of Educational Computing Research, 60(3), 645-670.

Corchón, E. (2000). La escuela rural: pasado, presente y perspectivas de futuro. Oikos-Tau.

Domínguez, I. (2010). Comunicación y texto. Editorial Pueblo y Educación.

Fernández, R. M. (2014). La superación profesional: una mirada desde la práctica educativa. Editorial Académica Española.

Freire, P. (1970). Pedagogía del oprimido. Siglo XXI Editores.

Fuego, M. R. (2003). La escuela rural y su relación con la comunidad. Revista Cubana de Educación Superior, 23(1), 45-60.

Guerra, S. (2014). Estrategias de superación profesional para docentes. Editorial Universitaria.

Guilarte, H. (2003). La formación del maestro para la escuela rural multigrado [Tesis de maestría no publicada]. Instituto Superior Pedagógico "Frank País García".

Handal, B., Watson, K., Petocz, P., & Maher, M. (2018). Choosing to teach in rural and remote schools: The zone of free movement. Education Research and Perspectives, 45, 1-25.

Herrera, J. L. (2024). Tendencias actuales en la superación profesional docente. Editorial Félix Varela.

Hockly, N. (2023). Artificial intelligence in English language teaching: The good, the bad and the ugly. RELC Journal, 54(2), 445-451.

Kearns, P., Knipe, S., & Foster, J. (2009). The impact of rurality on innovation and good practice in primary schools. Journal of Research in Rural Education, 24(7), 1-15.

Kessler, G. (2018). Technology and the future of language teaching. Foreign Language Annals, 51(1), 205-218.

López, Z. S., & Collazo, M. A. (2018). Enfoques teóricos acerca de la superación profesional, una mirada en las áreas técnicas. Revista Cubana de Educación Superior, 37(3), 134-148.

Ministerio de Educación Superior. (2018). Reglamento de la educación de posgrado de la República de Cuba. Gaceta Oficial de la República de Cuba.

Moussalli, S., & Cardoso, W. (2020). Intelligent personal assistants: can they understand and be understood by second language learners? Computer Assisted Language Learning, 33(8), 865-890.

Pokrivcakova, S. (2019). Preparing teachers for the application of AI-powered technologies in foreign language education. Journal of Language and Cultural Education, 7(3), 135-153.

Rayón, L., & De Las Heras, M. A. (2012). La relación escuela-comunidad en contextos rurales. Revista de Investigación Educativa, 30(1), 169-184.

Santos, J. (2005). La superación profesional: base para la calidad educativa. Editorial Pueblo y Educación.

Santos, J. (2010). Compendio de trabajo de posgrado para la Educación Técnica y Profesional. La formación permanente del profesor de la Educación Técnica y Profesional: La superación, el trabajo metodológico y la actividad científica como componentes principales (2da parte). Editorial Pueblo y Educación.

Silva Téllez, N., Alfonso Cruz, I., & Fernández Chelala, R. M. (2014). La superación profesional de los maestros primarios de la escuela rural. Revista Mendive, 12(4), 318-324.

Silva Téllez, N., Alfonso Cruz, I., & Fernández Chelala, R. M. (2022). Estrategia para la dirección del trabajo metodológico del maestro de la escuela multigrado. Revista Científica Hallazgos21, 7(3), 295-308. http://revistas.pucese.edu.ec/hallazgos21/

Valcárcel, N. (1998). La superación profesional: un reto para el docente del siglo XXI. Editorial Academia.

Valle, A. (1997). Teoría y práctica de la superación profesional de los docentes. Editorial Pueblo y Educación.

Conflicto de intereses

Los autores declaran no tener ningún conflicto de intereses.

Declaración de responsabilidad de autoría

Christian Elvis Villalta Pogo, Leonardo Mauricio Martínez Paredes, Francis Marcelo Pacheco Diaz y René Fabricio Villalta Pogo: revisión bibliográfica, investigación, corrección y redacción del artículo.