Teaching Technique Allessandra Elisabeth dos Santos, Research Group Vice-Leader: Technologies, Education, and Applied Linguistics (TECLA/CNPq), Universidade Federal de Sergipe, Brazil; email bethparis25@gmail.com. Optimizing AI Text Generators for Multilingual Learners: The Art of Crafting Effective Prompts GATESOL Journal 2025, Vol. 34(1), pp. 35–43 doi:10.52242/gatesol.193 ISSN: 2331-6845 Allessandra Elisabeth dos Santos Universidade Federal de Sergipe Abstract The transformative role of Generative Artificial Intelligence (GenAI) in language learning has shifted perspectives on the concept of writing. The challenges of writing are uniquely complex, encompassing a range of cognitive, cultural, and emotional dimensions. This article explores the crucial role of prompt engineering in leveraging the potential of GenAI platforms as a scaffolding strategy for writing with a focus on multilingual learners. By showcasing practical applications and examples, this teaching technique underscores the significance of well-structured prompts in providing scaffolding to support writing as a social practice, based on the Prompt Creation Reference Chart (dos Santos et al., 2025). Keywords Generative artificial intelligence (GenAI), prompt engineering, scaffolding, writing, multilingual learners Introduction Generative Artificial Intelligence (GenAI) occupies a prominent position in the contemporary landscape of technological advancements impacting societal contexts, including education. The transformative capacity of AI text generators in language learning has shifted perspectives on the concept of writing (dos Santos, 2024). The prior advent of digital technologies, which broadened the scope of literacy to encompass diverse multimodal forms – visual, auditory, and textual elements, has influenced this shift (Casanave, 2020). This expansion of literacy perspectives enriches learning experiences and challenges traditional educational frameworks. The challenges of writing are uniquely complex, encompassing a range of cognitive, cultural, and emotional dimensions (Casanave, 2018, 2020; Boa Sorte et al., 2021; Russell-Pinson & Harris, 2019). Specifically, writing anxiety impacts writing ability since it can affect the brainstorming of ideas and can lead to writer’s block (Wern & Rahmat, 2021). The introduction of AI-powered tools such as ChatGPT has significantly altered second language (L2) writing pedagogies by automating and enhancing the writing process, offering new methods for student engagement and learning (Yan, 2023). GenAI can be a powerful ally, providing tailored support through customized prompts that guide learners in navigating writing challenges (dos Santos et al., 2025). This article explores the crucial role of prompt engineering in leveraging GenAI platforms as a scaffolding strategy for writing, with a focus on multilingual learners (MLs). By showcasing practical applications and examples, this teaching technique underscores the significance of developing well-structured prompts to provide scaffolding to support writing as a social practice. mailto:bethparis25@gmail.com 36 GATESOL Journal 34(1) ChatGPT: Applications in Language Learning and Education AI text generators, of which ChatGPT is one example, can offer educators opportunities to develop higher levels of critical thinking by identifying and applying creative strategies for their integration (dos Santos et al., 2023). Educators can use ChatGPT to create study plans tailored to each student's needs, allowing for more effective and efficient learning (Skrabut, 2023). AI text generators1 can assist in generating questions that promote critical thinking and active engagement. GenAI platforms can help educators develop rubrics for assessing assignments, ensuring consistent grading. Teachers can employ ChatGPT to create realistic problem scenarios and case studies, providing students with practical applications of theoretical knowledge, thus preparing them for real-world challenges. ChatGPT: Ethics and Implications The utilization of GenAI in education presents opportunities and challenges. While students may benefit from improved writing efficiency, significant academic integrity concerns have surfaced (Jarrah et al., 2023; Yan, 2023). These include potential threats to traditional learning paradigms and the necessity to revisit plagiarism definitions in an AI-dominated era. Additionally, AI tools like ChatGPT pose ethical challenges in academia, especially in research and publishing. Reviewers and editors struggle to distinguish AI-generated content from original work, raising concerns about biases or misconduct like data fabrication (Casal & Kessler, 2023). Given these ethical challenges, it is important to discuss AI text generators’ impact on authorship, creativity, and plagiarism (Boa Sorte et al., 2021). There is a need to integrate clear guidelines into curricula to promote the ethical usage of AI tools, while acknowledging their limitations and biases. Such integration is crucial for maintaining educational value and fostering an environment that supports innovation and ethical responsibility (Yan, 2023). Practical Applications of AI Text Generators as a Scaffolding Strategy for Writing In writing instruction for MLs, scaffolding is crucial (de Oliveira & Jones, 2023). By building on learners’ existing knowledge and progressively increasing task complexity, scaffolding enhances L2 writing skills. Scaffolding includes guided practice, tailored feedback, and demonstrations of writing strategies through modeling, with explicit guidance gradually decreasing as learners' competence and confidence grow. When effectively employed, scaffolding can promote linguistic accuracy and integrate writing as a critical social practice. It equips learners to convey complex ideas and engage with diverse tasks across genres and disciplines, preparing them to navigate and contribute to various social contexts through skillful communication. Modern pedagogical frameworks can incorporate AI tools like ChatGPT and QuillBot, adapting traditional scaffolding to enhance interaction and personalized learning (Duncanson, 2024; Barrot, 2023). Teaching Example In this ninth-grade English Language Arts (ELA) class, students strengthen argument-based writing skills by producing 250 to 500-word book reviews. The students’ goal is to persuade peers to read the book they reviewed, part of their reading project. The task aligns with the ‘argue’ Key Language Use in the WIDA framework (WIDA, 2020, p. 219), requiring well-supported arguments with robust evidence and reasoning. This cognitive load can be challenging and may 1 “AI text generators” refers to a variety of text-based GenAI platforms (rather than other forms of GenAI such as music, images, or code). ChatGPT is one such example and is the platform used for the teaching technique illustrated. Dos Santos 37 trigger emotional reactions such as anguish and anxiety, especially when learners struggle with abstract thinking or articulating ideas (Casanave, 2018; Russell-Pinson & Harris, 2019). To address these cognitive and emotional challenges, learners are encouraged to collaborate in pairs or small groups, using ChatGPT as a resource. The collaboration is structured into five steps, described below. Step One: Initial Exploration of Prompt Writing and ChatGPT’s Responses. Students are encouraged to explore ChatGPT freely by writing prompts and discussing the outputs. For instance, one student’s prompt, “Write a review about the book [Title]”, lacks specific information (e.g., background, detailed characteristics of the task) and exemplifies unethical use of this platform since it bypasses the intellectual effort involved in writing, including critical analysis and creativity. Another student tried the prompt “How can I write a book review?” ChatGPT’s response was an extensive 14-step guide in bullet-point format, which can be overwhelming for some students because of the excessive amount of information. Nevertheless, the output provided general insights and some initial characteristics of a scaffolding strategy for writing. These human- AI interactions highlight the need for additional guidance to help students overcome writing challenges with ChatGPT. Step Two: Introducing the Prompt Creation Reference Chart. The teacher facilitates a discussion about the important role of prompt engineering in leveraging the potential of GenAI platforms. The teacher presents the Prompt Creation Reference Chart (dos Santos et al., 2025), a tool that enables learners to customize well-structured prompts to generate outputs that serve as scaffolding during writing. The Prompt Creation Reference Chart (see Figure 1) consists of three essential elements for the engineering of a well-structured, effective initial prompt: author, request, and purpose. The chart further describes each element as “(1) the prompt’s author: background information, (2) the main characteristics of the request: design, target audience, and theme, (3) and the purpose of the prompt: what the prompt is for, what the user wants to achieve” (dos Santos et al., 2025, p. 17). Step Three: Completing the Prompt Creation Reference Chart. Following the structure of the Prompt Creation Reference Chart, students can include the necessary descriptive information related to each of the three elements (see Figure 1). The teacher guides students through the chart completion process by eliciting information based on the format of the assignment. Step Four: Crafting the Prompt Following the Prompt Creation Reference Chart. With the importance of the teacher-to-student gradual release of responsibility in mind, the teacher modeled writing a ChatGPT prompt. The teacher used the chart as a guide, resulting in the following prompt (see Figure 2). 38 GATESOL Journal 34(1) Figure 1 Prompt Creation Reference Chart (Adapted from dos Santos et al., 2025) Figure 2 Creating a Customized Prompt Dos Santos 39 The selection of three colors aids in visually understanding prompt engineering. Upon analyzing the color-highlighted segments, it becomes apparent that the elements author and request are straightforward, boosting students’ confidence to begin writing their own prompts. However, the purpose element is pivotal in employing GenAI ethically as a scaffolding technique. At this point, educators should raise students’ awareness about ethical and purpose-driven questions directing the writing of prompts: “What is my prompt for?” and “What do I want to achieve with it?” These replace command-based prompts like “Write a review about the book [Title],” a misuse of GenAI in education settings, emphasizing the need for ethical modeling in human-AI interactions. Step Five: Discussing ChatGPT’s Response. Unlike the general guide created from the prompt “How can I write a book review?,” ChatGPT generated a structured set of instructions for composing a book review tailored for ninth-grade students. ChatGPT meticulously segments the review process into distinct parts, guiding students during the writing of the review. These include an introduction with a catchy title, a concise summary that avoids spoilers, character analysis, and themes (see Figure 3). Figure 3 ChatGPT Output (OpenAI, 2023) 40 GATESOL Journal 34(1) Dos Santos 41 Continuing with the review process, the guidance covers writing style, emotional impact, and a persuasive conclusion. Each section is equipped with precise word count limits and strategic advice, such as incorporating personal reflections, notable quotes, and comparisons to familiar works, enhancing both the engagement and analytical depth of the review. This structured guidance simplifies the writing task by encouraging critical thinking and effective communication among students. Conclusion This article demonstrates how GenAI can be used as a scaffolding strategy for MLs writing a persuasive book review assignment, and how prompt engineering can encourage students to engage critically with writing. Although planned with an ELA ninth-grade context in mind, the recommendations in this paper are transferable across other subjects, school grade levels, and higher education, supporting MLs as they navigate cognitive and emotional challenges in writing. The Prompt Creation Reference Chart (dos Santos et al., 2025) highlights key characteristics—author, request, and purpose—that guide the creation of structured prompts. These prompts scaffold MLs’ writing by breaking down the process into manageable steps, fostering ethical GenAI engagement, and enhancing AI response efficacy. Integrating GenAI supports various writing stages (Barrot, 2023; Duncanson, 2024; Yan, 2023), from idea generation to content structuring and grammar refinement, creating an adaptive and personalized learning environment. This approach maintains learner engagement through interactive feedback and fosters independence as GenAI gradually adjusts the support level based on learner progress. The adoption of structured prompts guides students in interacting with AI text generators, building their confidence and autonomy in writing. 42 GATESOL Journal 34(1) References Barrot, J. S. (2023). Using ChatGPT for second language writing: Pitfalls and potentials. Assessing Writing, 57, 100745. https://doi.org/10.1016/j.asw.2023.100745 Boa Sorte, P., Farias, M. A. de F., dos Santos, A. E., Santos, J. do C. A., & Dias, J. S. dos S. R. (2021). Artificial intelligence in academic writing: What is in store with the GPT-3 algorithm? Revista EntreLinguas, 7(00), 1–22. https://doi.org/10.29051/el.v7i00.15352 Casal, J. E., & Kessler, M. (2023). Can linguists distinguish between ChatGPT/AI and human writing?: A study of research ethics and academic publishing. Research Methods in Applied Linguistics. https://doi.org/10.1016/j.rmal.2023.100068 Casanave, C. P. (2018). Performing expertise in doctoral dissertations: Thoughts on a fundamental dilemma facing doctoral students and their supervisors. Journal of Second Language Writing, 43(4), 57–62. https://doi.org/10.1016/j.jslw.2018.02.005 Casanave, C. P. (2020). Controversies in second language learning: Dilemmas and decisions in research and instruction. The University of Michigan Press. de Oliveira, L. C., & Jones, L. (2023). Teaching young multilingual learners. Cambridge University Press. dos Santos, A. E., Boa Sorte, P., & de Oliveira, L. C. (2025). Generative artificial intelligence in writing: ChatGPT and critical questioning for multilingual learners. In C. Wang & Z. Tian (Eds.), Rethinking writing education in the age of generative AI (pp. 138–154). Routledge. dos Santos, A. E. (2024). Generative artificial intelligence and its impact on writing [Doctoral dissertation, Federal University of Sergipe]. https://ri.ufs.br/jspui/handle/riufs/20805 dos Santos, A. E., Olesova, L., Vicentini, C., & de Oliveira, L. C. (2023). ChatGPT in ELT: Writing affordances and activities. TESOL Connections. https://tcnewsletter.s3. amazonaws.com/newsmanager.commpartners.com/tesolc/issues/2023-05-01/2.html Duncanson, J. (2024). Supporting L2 writing using QuillBot. GATESOL Journal, 33(1), 33–39. https://doi.org/10.52242/gatesol.195 Jarrah, A. M., Wardat, Y. & Fidalgo, P. (2023). Using ChatGPT in academic writing is (not) a form of plagiarism: What does the literature say? Online Journal of Communication and Media Technologies, 13(4). https://doi.org/10.30935/ojcmt/13572 OpenAI. (2023). ChatGPT (Mar 23 version) [Large language model]. https://chat.openai.com/chat Russell-Pinson, L., & Harris, M. L. (2019). Anguish and anxiety, stress and strain: Attending to writers’ stress in the dissertation process. Journal of Second Language Writing, 43(1), 63– 71. https://doi.org/10.1016/j.jslw.2017.11.005 Skrabut, S. (2023). 80 ways to use ChatGPT in the classroom: Using AI to enhance teaching and learning [E-book]. Stan Skrabut. https://books2read.com/u/bQNV1w Wern, T., & Rahmat, N. (2021). An investigative study on the types and causes of ESL writing anxiety: A case study of learners from a Chinese independent middle school. European Journal of English Language Teaching, 6(3). http://dx.doi.org/10.46827/ejel.v6i3.3553 https://doi.org/10.1016/j.asw.2023.100745 https://doi.org/10.29051/el.v7i00.15352 https://doi.org/10.1016/j.rmal.2023.100068 https://doi.org/10.1016/j.jslw.2018.02.005 https://ri.ufs.br/jspui/handle/riufs/20805 https://tcnewsletter.s3.amazonaws.com/newsmanager.commpartners.com/tesolc/issues/2023-05-01/2.html https://tcnewsletter.s3.amazonaws.com/newsmanager.commpartners.com/tesolc/issues/2023-05-01/2.html https://doi.org/10.52242/gatesol.195 https://doi.org/10.30935/ojcmt/13572 https://chat.openai.com/chat https://doi.org/10.1016/j.jslw.2017.11.005 https://books2read.com/u/bQNV1w http://dx.doi.org/10.46827/ejel.v6i3.3553 Dos Santos 43 WIDA. (2020). WIDA English language development standards framework, 2020 edition: Kindergarten–grade 12. Board of Regents of the University of Wisconsin System. Yan, D. (2023). Impact of ChatGPT on learners in a L2 writing practicum: An exploratory investigation. Education and Information Technologies, 28(1), 13943–13967. https://doi.org/10.1007/s10639-023-11742-4 ____________________________________________________________________________ Received: February 2, 2024 | Accepted: May 2, 2025 | Published: October 2025 https://doi.org/10.1007/s10639-023-11742-4 GATESOL Journal Optimizing AI Text Generators for Multilingual Learners: The Art of Crafting Effective Prompts 2025, Vol. 34(1), pp. 35–43 doi:10.52242/gatesol.193 ISSN: 2331-6845