Using AI Simulated Conversations to Provide Speaking and Writing Practices
A Review of the Literature
By
Guadalupe Marin
Lamar University
February 15th, 2026
Introduction
Second language acquisition requires meaningful interaction, consistent practice, and opportunities for authentic communication (Masats, 2017; Soomro, 2022). In today’s global society, bilingualism provides increased professional opportunities and expanded cultural understanding. As a high school Spanish teacher, one of the greatest instructional challenges is the limited opportunity students have to interact with fluent or native Spanish speakers. While classroom instruction provides vocabulary and grammatical foundations, students often lack sustained conversational practice necessary to develop oral proficiency and communicative competence (Masats, 2017).
Research consistently emphasizes that speaking and writing are essential components of language development (Elkatmış, 2024; Soomro, 2022). However, when learners do not have sufficient opportunities to actively use the language, fluency decreases and vocabulary retention diminishes. Students frequently experience anxiety, shyness, and reduced confidence when asked to sustain real-time conversations (Nisrina, 2023). With the emergence of artificial intelligence (AI) tools capable of simulating natural language interactions, new opportunities exist to create structured, low-anxiety environments where students can practice conversational Spanish aligned with instructional goals (Piyumi et al., 2019; Üstünbas, 2024).
This literature review examines the role of AI-simulated conversational tools as a blended learning modality to support oral proficiency development (Horn & Staker, 2015). The review explores themes related to lack of authentic interaction, psychological barriers to speaking, AI-supported language learning environments, and the instructional advantages and limitations of artificial intelligence in education.
This body of research informs the following research question: How does the use of AI-simulated conversational tools impact high school Spanish students’ oral proficiency, confidence, and engagement in second language learning?
Review of the Literature
Definition of AI-Simulated Conversational Tools in Second Language Learning
AI-simulated conversational tools refer to large-scale language models capable of engaging users in natural language exchanges through text-based or interactive dialogue systems (Kos & Mažgon, 2025). Kos and Mažgon (2025) describe large language models as systems designed to process natural language tasks such as answering questions and problem-solving through human-like responses. Within educational contexts, these tools function as conversational agents that simulate back and forth communication, offering learners structured interaction without requiring a live conversational partner (Zulfikasari et al., 2024).
In second language learning, such tools can serve as digital conversational partners that allow students to practice vocabulary, grammar, and discourse in contextualized scenarios (Piyumi et al., 2019). Üstünbas (2024) highlights that AI chat systems can create interactive speaking practice opportunities that simulate conversational exchanges rather than passive exposure to content. These simulated interactions offer learners the opportunity to produce language in real time while receiving immediate responses, reinforcing the interactionist perspective that communication drives acquisition (Masats, 2017).
Types of AI-Supported Conversational Learning
AI Chat-Based Language Practice
AI chat platforms such as ChatGPT utilize natural language processing to generate human-like dialogue that supports language practice (Zulfikasari et al., 2024). These systems enable learners to engage in structured conversations aligned to specific prompts, tasks, or themes. Tarp and Nomdedeu-Rull (2024) note that AI-driven language models have already been used to support Spanish writing assistance, demonstrating their capacity to scaffold language production and provide corrective feedback. Research on synchronous computer mediated feedback further suggests that digital conversational exchanges can enhance language development through interactive correction and response (Piyumi et al., 2019).
AI as Instructional Support Tools
Beyond conversational practice, artificial intelligence tools function as academic assistants that provide rapid feedback and personalized support. Hanshaw et al. (2024) found that AI course assistants can enhance academic achievement and improve self-efficacy by offering immediate responses and encouraging student participation. Similarly, Jumriah et al. (2024) emphasize that AI tools can enhance student performance by facilitating comprehension and engagement across subject areas. These tools extend beyond simple automation and function as interactive systems capable of guiding learners through structured tasks while increasing motivation and persistence (Hanshaw et al., 2024).
Blended Integration in Educational Settings
Educational institutions are increasingly proposing the integration of AI tools in teaching, learning, and assessment processes to enhance instructional effectiveness (Onuoha et al., 2024). In blended environments, AI tools supplement teacher-led instruction by providing flexible practice opportunities outside of scheduled class time (Horn & Staker, 2015). This modality allows students to continue practicing language skills beyond the classroom setting while maintaining alignment with curriculum objectives and communicative goals (Masats, 2017).
Advantages of Using AI-Simulated Conversational Tools
One major advantage of AI-supported conversational practice is increased access to interactive opportunities. A significant barrier in second language classrooms is the lack of authentic conversational partners (Mrak, 2020). Üstünbas (2024) explains that effective language learning requires engagement in back-and-forth interaction rather than passive observation. AI-generated conversations can provide structured opportunities for this type of engagement, particularly in environments where access to native speakers is limited, reinforcing communicative competence development (Masats, 2017).
Another advantage is the reduction of psychological barriers associated with speaking. Research identifies anxiety, shyness, and confidence as major factors affecting speaking performance (Nisrina, 2023). Soomro (2022) also found that insufficient speaking practice increases learner anxiety when students attempt to sustain conversations, a finding similarly supported by Nisrina (2023). Research on computer-mediated communication suggests that digital environments can provide flexible, low-anxiety settings that support language development (Piyumi et al., 2019). By offering immediate feedback and flexible pacing, AI tools may support increased confidence and intrinsic motivation (Hanshaw et al., 2024).
Additionally, AI tools can extend learning beyond the classroom. When heritage or second languages are not dominant in students’ communities, opportunities for authentic communication are limited (Mrak, 2020). AI platforms can provide consistent exposure and practice in contexts where cultural immersion is not readily available. This accessibility aligns with blended learning principles that emphasize flexibility and learner autonomy (Horn & Staker, 2015), supporting continued skill development and vocabulary retention.
Barriers to Implementing AI-Simulated Conversational Tools
Despite their potential benefits, AI tools present several challenges. Soomro (2022) notes that acquiring a second language is a complex process influenced by multiple contextual and instructional factors. While AI tools may simulate conversation, they cannot fully replicate the cultural nuance and spontaneity of human interaction emphasized in communicative language teaching (Masats, 2017).
There are also concerns regarding over reliance on technology. Onuoha et al. (2024) acknowledge both the benefits and challenges of AI integration, emphasizing the need to balance traditional pedagogical approaches with emerging technologies. Kos and Mažgon (2025) further caution that large language models must be implemented thoughtfully to avoid replacing essential instructional practices and authentic interaction opportunities.
Finally, although AI tools may reduce anxiety during digital interaction, it remains necessary to examine whether this confidence transfers to peer-to-peer or real-world conversations. Speaking proficiency requires not only comfort but also adaptability, comprehension, and spontaneous response skills that develop through authentic interaction (Masats, 2017).
Psychological and Linguistic Factors in Speaking Development
Language production requires complex cognitive and emotional engagement (Soomro, 2022). Elkatmış (2024) emphasizes that writing and expressive language tasks require extensive vocabulary, critical thinking, and consistent practice. Similarly, anxiety significantly impacts speaking performance (Nisrina, 2023). The development of oral proficiency depends not only on exposure to language but also on opportunities for active production in supportive environments (Piyumi et al., 2019).
Olasik (2023) highlights the relational aspect of interacting with AI systems, noting that users often perceive conversational agents as expressive and responsive. This perception may contribute to learner engagement and sustained participation. If AI systems can foster a sense of conversational continuity, they may provide learners with additional opportunities to rehearse language skills in a psychologically safe setting, supporting communicative growth (Masats, 2017).
Summary
The literature demonstrates that second language learners face significant challenges related to limited authentic interaction, psychological barriers, and insufficient practice opportunities (Mrak, 2020; Nisrina, 2023). AI-simulated conversational tools have emerged as interactive systems capable of generating human-like dialogue that may support language production, confidence, and extended learning opportunities (Zulfikasari et al., 2024; Üstünbas, 2024). However, the research also emphasizes the importance of maintaining balance between technological innovation and traditional pedagogical practices grounded in communicative interaction (Horn & Staker, 2015; Masats, 2017).
This Review and the Field of Education
This body of literature contributes to the broader field of education by examining how emerging artificial intelligence technologies can be integrated into language instruction to address persistent challenges such as anxiety, limited interaction, and access disparities (Onuoha et al., 2024). As schools navigate AI integration, understanding how these tools can function as supplemental instructional supports rather than replacements for teacher-led instruction is critical (Kos & Mažgon, 2025). For World Language departments, AI-simulated conversation may provide scalable opportunities to enhance oral practice while maintaining curriculum alignment and communicative competence development (Masats, 2017).
Strengths and Weaknesses of This Body of Literature
One strength of the literature is the consistent recognition of interaction and practice as central to language development (Masats, 2017; Soomro, 2022). Multiple studies support the role of AI tools in increasing engagement, self-efficacy, and academic performance (Hanshaw et al., 2024; Jumriah et al., 2024). Additionally, research highlights the psychological dimensions of speaking, reinforcing the importance of reducing anxiety in language classrooms (Nisrina, 2023; Ümran, 2024).
However, limitations exist. Much of the current literature focuses on higher education or generalized AI implementation rather than K–12 secondary language classrooms (Elkatmış, 2024; Tarp & Nomdedeu-Rull, 2024). There is limited empirical research specifically examining AI-simulated conversational tools and measurable growth in oral proficiency. Furthermore, few studies investigate whether improvements in AI-supported practice transfer to improved peer-to-peer communication grounded in authentic interaction (Masats, 2017).
Focus of the Current Study
The current study addresses these gaps by implementing AI-simulated conversational tools within a high school Spanish classroom using a mixed-methods action research design. By collecting pre- and post-speaking assessments alongside student reflections and observational data, this research seeks to examine in what ways AI-supported conversation impacts oral proficiency, confidence, and engagement. This classroom-based investigation will provide practical insights for educators considering AI integration while maintaining emphasis on communicative competence (Masats, 2017) and blended instructional practice (Horn & Staker, 2015).
References
Elkatmış, M. (2024). ChatGPT and creative writing: Experiences of master’s students in enhancing writing skills. International Journal of Contemporary Educational Research, 11(3). https://doi.org/10.52380/ijcer.2024.11.3.597
Hanshaw, G., Vance, J., & Brewer, C. (2024). Exploring the effectiveness of AI course assistants on the student learning experience. Open Praxis, 16(4), 627–644.
Horn, M. B., & Staker, H. (2015). Blended: Using disruptive innovation to improve schools. Jossey-Bass.
Jumriah, S. E. S., Supriatna, E., Smas, M. H., & Arini, I. (2024). Analysis of the use of ChatGPT to improve student performance. Edukasi, 5(1).
Kos, Z., & Mažgon, J. (2025). The challenges of using large language models: Balancing traditional learning methods with new technologies in the pedagogy of sociology. Education Sciences, 15(2). https://doi.org/10.3390/educsci15020191
Masats, D. (2017). Conversation analysis at the service of research in the field of second language acquisition (ERIC No. ED573594). ERIC. https://eric.ed.gov/?id=ED573594
Mrak, A. (2020). Developing writing in Spanish heritage language learners: An integrated process approach. Dimension, 82–96.
Nisrina, N. (2023). Factors that affect speaking skills of students from ethnic minorities in English language learning. Bahasa Dan Seni: Jurnal Bahasa,
Sastra, Seni, Dan Pengajarannya, 51(1), 120–132. https://doi.org/10.17977/um015v51i12023p120
Onuoha, I., Okocha, E. I., Aneke, O., Adekunle, O. E., & Anthony, U. (2024). Benefits and challenges of utilizing ChatGPT and Gamma.app in teaching, learning, and assessment of vocational students in 21st century colleges of education in Enugu State. Proceedings of the Nigerian Academy of Science, 17(2), 34–45. https://doi.org/10.57046/RGDY2454
Olasik, M. (2023). “Good morning, ChatGPT, can we become friends?” An interdisciplinary scholar’s experience of getting acquainted with OpenAI’s ChatGPT: An autoethnographical report. European Research Studies, 26(2), 269–284. https://doi.org/10.35808/ersj/3168
Piyumi, W. A., Ola, K., & Sirkku, M. B. (2019). Complexity and potential of synchronous computer-mediated corrective feedback. In Proceedings of the EUROCALL 2019 Conference. Research-publishing.net.
Soomro, A. S. (2022). Challenges faced by second language learners: A review. Sindh Journal of Linguistics, 1(1), 51–57. https://doi.org/10.58921/sindhjol.v1i1.10
Tarp, S., & Nomdedeu-Rull, A. (2024). Who has the last word? Lessons from using ChatGPT to develop an AI-based Spanish writing assistant. Círculo de Lingüística Aplicada a la Comunicación, 97. https://doi.org/10.5209/clac.91985
Üstünbas, U. (2024). Hey, GPT, can we have a chat? A case study on EFL learners’ AI speaking practice. International Journal of Modern Education Studies, 8(1), 91–107.
Zulfikasari, S., Sulistio, B., & Aprilianasari, W. (2024). Utilization of ChatGPT artificial intelligence (AI) in students’ learning experience in Gen-Z classrooms. Lectura: Jurnal Pendidikan, 15(1), 259–272. https://doi.org/10.31849/lectura.v15i1.18840