Using GenAI for Feedback and Growth in TESOL Multimodal Writing

Abstract

This article explores the pedagogical integration of Generative AI (GenAI) tools in TESOL writing classrooms, with a focus on digital multimodal composing (DMC). It examines how AI can support formative feedback, rubric-based assessment, and language scaffolding, while also addressing concerns around authenticity and assessment validity. Emphasizing ethical implementation, teacher agency, and learner development, the article offers practical recommendations for aligning GenAI with second language acquisition (SLA) principles. Through structured tasks, reflective practices, and responsible AI use, educators can enhance writing instruction and promote effective multimodal communication in English language learning

Introduction

The rapid evolution of artificial intelligence in recent years has opened new possibilities for teaching and learning across disciplines. In the field of TESOL (Teaching English to Speakers of Other Languages), one particularly promising development is the integration of Generative AI (GenAI) into multimodal writing instruction. As language classrooms become increasingly digital, TESOL educators are exploring how tools like ChatGPT, Claude, and other AI-driven platforms can support student growth, particularly in the areas of formative feedback and dynamic assessment. This article aims to investigate how GenAI can be utilized as a meaningful and pedagogically sound component in the process of digital multimodal composing (DMC).

Multimodal writing tasks go beyond traditional essays and written texts. They encompass a range of communicative forms such as video presentations, digital storytelling, infographics, and mixed media projects. These assignments allow English language learners to engage in more expressive and personalized modes of communication, integrating linguistic, visual, and auditory elements. However, this also places new demands on instructors, who must assess not only language accuracy and coherence but also visual design, interactivity, and overall communicative effectiveness. GenAI tools have emerged as potential allies in managing these complex dimensions by offering real-time feedback, supporting peer review, and contributing to rubric-based assessment processes.

Despite the promising capabilities of GenAI, its integration into TESOL writing pedagogy must be approached with caution and care. Concerns regarding learner authenticity, over-reliance on automated output, and the ethical use of AI in assessment continue to shape this conversation. Educators must remain attentive to the risks of allowing AI tools to replace rather than support the teacher’s professional judgment and the learner’s developmental process. The goal is not to automate feedback entirely, but to enrich the feedback ecosystem and promote metacognitive engagement through accessible, student-friendly interventions.

This article is intended for a broad audience of TESOL educators, including classroom teachers, curriculum developers, assessment specialists, and teacher-researchers who are interested in incorporating AI tools into writing instruction. While the focus remains grounded in professional practice, the discussion will balance academic rigor with readability, offering clear explanations and practical applications. Each section will explore a key aspect of using GenAI in multimodal writing, from theoretical foundations and feedback design to rubric development and concerns about authenticity.

The sections that follow will provide a comprehensive framework for integrating GenAI in multimodal writing pedagogy. By examining the pedagogical potential and limitations of these tools, the article aims to contribute to informed, respectful, and effective practices in English language teaching and assessment.

Understanding TESOL Multimodal Writing in the Digital Age

Defining digital multimodal composing in TESOL

In recent years, the concept of writing in the TESOL context has undergone a significant transformation. No longer confined to linear, alphabetic text, writing now encompasses a broad range of communicative modes. This change has given rise to the practice of digital multimodal composing (DMC), wherein learners integrate visual, auditory, spatial, and textual elements to create meaning. These compositions may take the form of video essays, digital narratives, interactive presentations, or infographics, allowing learners to convey ideas through multiple semiotic channels.

Digital multimodal composing is grounded in the understanding that communication is not limited to language alone. The New London Group (1996) first articulated this idea in their framework of multiliteracies, arguing for pedagogical approaches that reflect the dynamic and multimodal nature of meaning-making in contemporary contexts. In TESOL, this shift enables learners to demonstrate language proficiency through varied modes, fostering creativity while accommodating different communicative strengths.

Hafner (2021) defines DMC as “the process of creating texts that combine modes such as writing, images, video, and audio using digital technologies” (p. 237). For English language learners, this approach provides opportunities to engage with language in ways that more closely mirror real-world communication. It also allows learners to draw upon multiple resources to make meaning, rather than relying solely on linguistic accuracy.

How multimodality shifts writing pedagogy

The integration of multimodality into writing pedagogy requires educators to rethink both instructional strategies and assessment criteria. Traditional writing instruction typically focuses on sentence structure, grammar, cohesion, and clarity of argument. However, multimodal writing demands attention to layout, visual design, multimedia integration, and user interaction. These dimensions introduce new literacies that are essential for academic and professional communication in digital environments (Smith & Craig, 2020).

From a pedagogical standpoint, the shift to DMC involves balancing linguistic goals with broader communicative objectives. Ortega (2017) emphasizes that teaching multimodal writing in TESOL settings should not come at the expense of language development, but rather should be used to enhance it. When guided effectively, multimodal projects can scaffold language learning by providing meaningful contexts for vocabulary use, grammar practice, and genre awareness.

For instance, when students produce a narrated video or a digital story, they must script their narration (supporting writing fluency), record spoken text (developing pronunciation), and select visual components (enhancing critical thinking and coherence). These interwoven tasks offer a more holistic form of language learning that supports both receptive and productive skills.

Moreover, the instructional shift to multimodality encourages a re-examination of process-oriented approaches. Writing becomes more iterative, involving planning, drafting, editing, and redesigning across modes. Teachers are called to provide targeted feedback not only on textual content but also on visual clarity, alignment between audio and text, and the overall rhetorical effect of a multimodal product (Hafner, 2021).

Student needs in multimodal writing environments

TESOL students engaging in multimodal writing face new challenges that differ from those encountered in traditional writing tasks. These include navigating digital platforms, understanding visual literacy principles, and coordinating different modes of communication into a coherent whole. Students also require guidance on the rhetorical choices that influence how different audiences interpret their multimodal compositions.

Research indicates that many learners benefit from explicit instruction in how to analyze and produce multimodal texts. Smith and Craig (2020) argue that without structured support, students may struggle to move beyond superficial use of images or video, failing to achieve effective communication. This underscores the importance of pedagogical scaffolding that integrates both digital literacy and language development.

Equally important is the need for accessible tools and resources. Students vary in their prior experience with multimedia tools, which can affect their confidence and performance. Teachers can support learners by selecting user-friendly platforms and providing step-by-step guides or tutorials. Providing model texts and exemplars is another strategy that helps learners understand expectations and develop their own voice in multimodal contexts (Hafner & Ho, 2020).

Assessment practices must also evolve to meet the changing needs of students. Formative assessment plays a particularly critical role in multimodal writing, offering learners ongoing feedback that guides both linguistic and multimodal development. Effective formative feedback should address the integration of modes, the alignment of language with visual or audio elements, and the communicative effectiveness of the overall composition.

Ortega (2017) further notes the importance of respecting the learner’s intentions in multimodal writing. Teachers should aim to understand what learners are attempting to convey and offer feedback that supports their communicative goals, rather than enforcing rigid expectations based solely on native speaker norms. This respectful approach to feedback encourages learner autonomy and fosters meaningful engagement with the writing process.

In summary, the emergence of digital multimodal composing in TESOL reflects broader changes in how communication is conceived and practiced. As students engage in more complex and authentic writing tasks, educators must develop new strategies for instruction, feedback, and assessment. When thoughtfully implemented, multimodal writing supports a more comprehensive approach to language learning that is both creative and academically grounded.

The Role of GenAI in TESOL Writing Classrooms

Overview of Generative AI capabilities relevant to writing

Generative Artificial Intelligence (GenAI) has gained considerable attention in education for its capacity to support writing instruction through real-time language generation, feedback, and scaffolded task support. In TESOL writing classrooms, tools such as ChatGPT, Claude, Gemini, and others are increasingly being explored for their pedagogical utility. These models are trained on vast corpora of language data and can produce coherent, contextually relevant text in response to user prompts, offering both native-like language samples and tailored responses to specific writing tasks.

Key capabilities of GenAI relevant to TESOL writing include the generation of model texts, assistance with brainstorming and outlining, reformulation of sentences for grammatical accuracy or stylistic improvement, and provision of automated feedback on drafts. Learners can iteratively interact with these tools, enabling a dynamic and responsive writing process. For instance, students may use GenAI to rephrase ideas for clarity, expand their vocabulary, or test different genre conventions in academic or creative writing formats (Kohnke & Moorhouse, 2023).

GenAI tools also offer low-stakes environments for language experimentation. Because these platforms provide immediate, private feedback, students may feel more confident in testing linguistic structures or expressing ideas without fear of public correction. In multilingual classrooms, GenAI can also be configured to translate or compare linguistic structures, assisting learners in transferring knowledge from their first language to English (Lu, 2023).

AI and second language acquisition principles

To effectively incorporate GenAI in TESOL writing instruction, it is essential to align its use with established principles of second language acquisition (SLA). The interactionist perspective in SLA emphasizes meaningful communication, negotiation of meaning, and the role of feedback in interlanguage development. GenAI tools, when used interactively, can simulate some aspects of this communicative process, offering language input and reformulations that resemble features of negotiated interaction.

Moreover, GenAI can support the noticing hypothesis (Schmidt, 1990), which posits that learners must consciously notice linguistic features in order to acquire them. By highlighting grammatical issues or offering alternative phrasings, GenAI can help learners become aware of gaps in their linguistic knowledge. This aligns with Elgort’s (2022) observation that learners benefit most when AI-supported writing tools are integrated into task-based learning, where language form is addressed in the context of meaningful output.

In terms of scaffolding, GenAI can function as a form of assisted performance within Vygotsky’s Zone of Proximal Development (ZPD). Learners engage with language that is slightly above their current level and receive feedback that guides them toward higher proficiency. Godwin-Jones (2023) suggests that this form of digital mediation supports learner autonomy while reinforcing language development goals, especially when teachers model responsible and reflective AI use.

However, to ensure alignment with SLA goals, GenAI use must be framed as a support mechanism rather than a replacement for learner effort. It should facilitate, not bypass, the cognitive and communicative processes that underpin language acquisition.

Concerns about over-reliance and misuse

Despite its potential, the integration of GenAI in TESOL writing classrooms raises several pedagogical and ethical concerns. One of the most frequently cited issues is the risk of over-reliance. When students begin to depend heavily on GenAI tools to generate, revise, or even complete writing tasks, there is a danger that the tool supplants the learner’s own language processing and critical thinking. This undermines the developmental goals of writing instruction and may result in superficial learning.

Kohnke and Moorhouse (2023) caution that while GenAI can help learners produce grammatically correct and fluent texts, these outputs may not always reflect a learner’s true proficiency. As a result, instructors may find it difficult to assess students’ actual abilities or progress over time. Misuse of AI can therefore lead to a distorted understanding of learner development and weaken the validity of both formative and summative assessments.

Furthermore, there are important ethical considerations related to authorship and authenticity. In educational settings, it is essential to distinguish between appropriate AI assistance and unacknowledged substitution. Lu (2023) emphasizes the need for transparent guidelines that help students understand when and how to use GenAI tools responsibly. Without such guidance, there is a risk that students may unknowingly engage in academic dishonesty by submitting AI-generated content as their own.

The use of GenAI also raises questions about data privacy, accessibility, and content accuracy. Although GenAI tools are designed to generate coherent language, they can sometimes produce factually incorrect or culturally insensitive content. TESOL educators must therefore act as critical mediators, helping learners verify, adapt, and reflect on the AI-generated output. Teachers should also stay informed about the terms of service and data usage policies associated with these tools to protect student privacy.

In response to these concerns, a balanced pedagogical approach is required—one that harnesses the strengths of GenAI while fostering learner agency and ethical awareness. Teachers should explicitly teach students how to critically engage with AI-generated feedback, revise based on their own goals, and retain ownership of their writing process. Classroom policies should promote transparency, emphasize learning over perfection, and incorporate reflective activities that ask students to document and evaluate their use of AI tools.

In summary, GenAI holds promise as a valuable resource in TESOL writing classrooms, offering learners dynamic language support and enabling teachers to enhance feedback practices. When grounded in SLA theory and guided by thoughtful implementation, these tools can contribute to writing development without compromising authenticity or instructional integrity. However, to maximize their pedagogical benefit, educators must remain vigilant about the risks of overuse and misuse and proactively establish frameworks that uphold both academic and learner-centered values.

Using GenAI to Support Feedback in Multimodal Writing

Automated and human-in-the-loop feedback

As digital multimodal composing (DMC) becomes a standard part of TESOL instruction, the need for effective feedback systems has grown. In this context, Generative AI (GenAI) tools can play a valuable role by supporting both automated and hybrid feedback models. Automated feedback refers to real-time, AI-generated responses that provide learners with immediate input on their written work, while human-in-the-loop feedback combines the speed and scalability of AI with the nuance and pedagogical judgment of human educators.

Automated feedback has proven useful for certain aspects of writing, such as grammar correction, vocabulary enhancement, and sentence clarity. Tools like Grammarly, ChatGPT, and QuillBot can identify issues in mechanics and syntax and suggest revisions accordingly. For TESOL learners, these tools reduce wait time and provide a low-pressure environment to revise and improve their drafts (Ranalli, 2021). In DMC tasks, GenAI can also assist in evaluating caption-text alignment, sentence cohesion in audio narration scripts, and the clarity of accompanying textual descriptions.

However, AI alone is insufficient for the deeper interpretive work involved in assessing multimodal compositions. Human-in-the-loop feedback ensures that the AI-generated suggestions are filtered through pedagogical principles. Teachers can review, modify, and expand upon the automated input, maintaining control over the learning outcomes while using AI to enhance efficiency. Wang et al. (2023) emphasize that combining AI with teacher-led commentary enables more holistic feedback, particularly when visual, auditory, and textual elements must be evaluated together.

The strength of this hybrid approach lies in its responsiveness. While GenAI tools deliver immediate, objective data, teachers provide contextual, formative guidance. For instance, an AI tool might flag a sentence as awkward, but it cannot always discern whether the phrasing was part of a rhetorical strategy or the result of language transfer. A teacher, understanding the student’s intent and linguistic background, can interpret the feedback appropriately. In DMC tasks, this layered model allows instructors to support both language learning and design literacy, creating a feedback loop that is both timely and pedagogically sound (Warschauer & Liaw, 2023).

How AI can support peer and self-review

In addition to teacher-led feedback, peer and self-assessment are vital components of writing development. These practices encourage learners to reflect on their own choices and become more independent editors and evaluators. GenAI tools can scaffold these processes by offering prompts, suggestions, and models that guide learners through the review process.

For example, learners can input their own drafts into a GenAI interface and receive automated summaries of strengths and weaknesses, which they can then reflect on during self-assessment. This may include insights into tone, cohesion, vocabulary usage, or even genre alignment. Such feedback helps learners internalize language norms and identify areas for revision without always requiring teacher intervention (Lee, 2022).

When integrated into peer review sessions, GenAI can offer sentence-level or structural suggestions that peers might overlook. By allowing students to verify each other’s observations using AI tools, peer assessment becomes more evidence-based. Moreover, learners can use AI-generated questions or checklists to guide their review, helping them to engage more critically with multimodal features such as layout consistency, image-text coherence, and narrative sequencing.

Warschauer and Liaw (2023) point out that peer interactions mediated through AI can promote language awareness, especially when students are encouraged to compare different revision suggestions and justify their choices. Rather than diminishing student agency, GenAI tools can empower learners to become more informed participants in the feedback process. With appropriate training, students can learn to critically evaluate AI suggestions, integrate relevant insights, and reject inappropriate ones.

Educators play a key role in modeling this evaluative behavior. Teachers can demonstrate how to interpret AI-generated comments, revise based on constructive feedback, and maintain their original voice and intent. Reflection logs and revision memos can further deepen metacognitive engagement by prompting students to explain their choices and document how they responded to AI and peer feedback.

Feedback on linguistic and visual-textual integration

Providing effective feedback on multimodal compositions requires more than pointing out grammar errors or improving vocabulary. Teachers must also address the alignment between linguistic and visual elements, the clarity of multimodal communication, and the overall effectiveness of message delivery. GenAI can assist in this area by evaluating coherence across modes, suggesting alternative phrasings for multimedia captions, and checking the semantic consistency between text and imagery.

For instance, if a student creates a video presentation with spoken narration and on-screen text, GenAI can analyze both scripts and images to assess redundancy, relevance, and support for the main message. When combined with design principles, this analysis can enhance the communicative quality of the product. AI can also evaluate subtitles or captions for conciseness, grammatical accuracy, and timing with audio tracks, supporting learners in creating professional-quality media.

Wang et al. (2023) highlight that GenAI can help bridge the gap between language and multimodal elements by guiding learners to create more coherent and polished compositions. However, interpretation still requires human sensitivity. The meaning of a visual or audio element is often culturally or contextually specific. While AI can evaluate technical aspects, only human reviewers can assess whether the integration of elements contributes to meaningful communication.

Teachers can guide students in evaluating visual-textual cohesion by using AI as a springboard for discussion. For example, students might compare AI-generated suggestions with their own interpretations, refining their multimodal choices based on audience expectations and communicative goals. In doing so, students learn to see GenAI not as an authority, but as a resource that prompts deeper inquiry into how meaning is constructed.

Moreover, feedback in multimodal writing should be cyclical and dialogic. AI tools provide the initial input, but it is the dialogue between student, peer, and teacher that fosters real growth. Effective feedback strategies involve guiding students to make informed decisions, experiment with alternatives, and reflect on their choices. The goal is not simply to correct errors but to support the development of confident, reflective communicators.

GenAI offers TESOL educators a practical tool for enhancing feedback in multimodal writing tasks. Automated and hybrid feedback systems can provide scalable, consistent input while maintaining the professional judgment of teachers. These systems, when thoughtfully applied, can also support peer and self-review, helping learners develop greater independence and critical engagement with their work.

However, the success of GenAI in the feedback process depends on careful integration. Teachers must model how to interpret, apply, and critique AI-generated suggestions, fostering a feedback culture that values process over perfection. In multimodal writing, where the interplay of visual and linguistic elements is central, GenAI can assist but not replace the human insight needed to evaluate meaning and effectiveness.

By embedding GenAI tools within pedagogically sound frameworks, educators can leverage technology to support student growth, enhance feedback cycles, and develop more responsive writing instruction.

Designing AI-Supported Rubrics for Multimodal Assessment

Multimodal writing rubrics and AI alignment

The assessment of digital multimodal composing (DMC) presents unique challenges in TESOL classrooms. Unlike traditional essays, multimodal projects require the integration of visual, textual, and sometimes interactive elements. This complexity necessitates the use of rubrics that can evaluate not only language use but also the coherence, design, and communicative purpose of multiple semiotic modes. When thoughtfully designed, these rubrics can align with the capabilities of Generative AI (GenAI) tools, allowing educators to enhance feedback quality and consistency across learner submissions.

Rubrics for multimodal writing need to encompass criteria that reflect the expanded nature of digital literacy. Hafner and Ho (2020) propose multimodal assessment frameworks that evaluate content accuracy, organization, use of digital tools, audience awareness, and multimodal orchestration. For TESOL learners, such rubrics offer a structured way to understand performance expectations across both linguistic and non-linguistic dimensions.

Aligning rubric categories with GenAI functionalities allows educators to efficiently distribute feedback efforts between automated and human assessments. For example, GenAI can evaluate textual elements for grammar, cohesion, and vocabulary variety, while teachers concentrate on visual rhetoric, design aesthetics, and alignment with task objectives. This dual approach not only saves time but also enhances the depth and breadth of feedback provided.

Moreover, GenAI can assist in standardizing rubric application. When integrated into digital learning platforms, AI models can analyze student submissions using predefined criteria, flagging areas of concern or excellence according to rubric descriptors. Lim (2022) suggests that this alignment can help reduce subjective variation in teacher judgments, especially in large or asynchronous classrooms. However, the rubric design must be detailed and precise enough to guide AI analysis and interpretation accurately.

Integrating AI for formative rubric-based assessment

Rubrics are particularly valuable in formative assessment, where the goal is to guide learning rather than merely assign scores. GenAI tools can support this formative function by delivering rubric-based feedback in real time, helping students identify which aspects of their multimodal texts require improvement. When students receive timely, rubric-aligned feedback, they are better positioned to revise and reflect on their work before final submission.

In practice, AI-assisted formative assessment can be operationalized by creating digital rubrics that integrate with AI systems. For instance, an AI tool may highlight sentence-level cohesion issues under a “textual coherence” rubric category or suggest improvements for visual consistency under a “layout and design” criterion. Coniam (2021) emphasizes that such systems can encourage learners to engage more actively with the rubric during the drafting process, thereby reinforcing learning goals and task clarity.

Additionally, GenAI can support iterative feedback cycles, where students upload revised drafts in response to initial rubric-based suggestions. This enables educators to track development over time and encourages students to approach writing as a process rather than a single event. When used in this way, GenAI becomes a feedback partner that facilitates progression through guided, criteria-focused input.

Educators can further enhance this process by involving students in rubric interpretation. Having learners compare AI-generated feedback with rubric descriptors fosters rubric literacy and increases student investment in their own learning. When students are taught how to navigate rubrics and reflect on AI feedback in relation to those expectations, they gain a clearer understanding of both the standards and their progress toward meeting them.

Ensuring transparent and respectful evaluation

While GenAI offers potential advantages in applying rubrics consistently and efficiently, it also raises questions about transparency and learner trust. To ensure that AI-influenced assessment is perceived as fair and respectful, educators must make the process visible and understandable to students. This involves explicitly explaining how the AI tool interprets rubric criteria and how its feedback should be used in conjunction with teacher guidance.

Ajjawi and Boud (2017) argue that effective feedback should be dialogic, enabling students to make sense of evaluative information and act upon it. This principle applies equally to AI-assisted feedback. Students need to understand that AI suggestions are not prescriptive judgments but rather cues for consideration. Teachers should encourage learners to ask questions, compare AI feedback with their own interpretations, and retain agency in the revision process.

Transparency also requires that the limits of AI evaluation be clearly communicated. GenAI tools, while capable of analyzing language and certain design features, are not equipped to assess subjective dimensions such as creativity, originality, or cultural relevance. Teachers must contextualize AI feedback within broader pedagogical goals, ensuring that learners view the AI tool as a complement rather than a replacement for human evaluation.

To support respectful assessment practices, institutions should provide guidelines for responsible GenAI use, addressing issues such as authorship, data privacy, and learner autonomy. Educators can reinforce these guidelines by modeling ethical practices in their own use of AI and by creating assessment environments where student voice and judgment are central to the process.

Another important aspect of respectful evaluation involves avoiding over-reliance on AI judgment. While GenAI can apply rubric criteria consistently, its feedback must be filtered through the professional insights of teachers who understand learner histories, needs, and intentions. Human evaluation remains essential for interpreting meaning, assessing interpersonal nuance, and fostering rapport between teacher and learner.

Ultimately, rubrics in multimodal writing are not only assessment tools but also teaching instruments. When aligned with GenAI, they become dynamic frameworks that support real-time, constructive engagement with writing tasks. By designing transparent, well-structured rubrics and integrating AI in ways that respect learner agency and instructional integrity, TESOL educators can offer more responsive, fair, and pedagogically sound feedback.

Scaffolding Language and Content Development with GenAI

Supporting genre awareness and structure

A fundamental component of writing development in TESOL classrooms is genre awareness. Understanding how different types of texts are structured and the communicative purposes they serve is essential for English language learners, especially when engaging in digital multimodal composing (DMC). Generative AI (GenAI) tools offer useful support by modeling genre-specific features, providing outlines, and suggesting appropriate language for various contexts.

GenAI systems such as ChatGPT can generate example texts in a range of genres—narrative, argumentative, descriptive, or expository—based on a brief prompt. When used thoughtfully, these tools help learners notice key structural elements of genres, such as thesis statements in opinion pieces or sequencing language in process descriptions. By prompting AI to produce sample texts that reflect academic or real-world genres, students can analyze and internalize patterns in organization and language use (Reinhardt, 2020).

Furthermore, GenAI can scaffold learners’ initial planning and outlining stages. For example, students might input a topic and receive a basic structure or outline, which they can then customize. This can be particularly valuable for learners with limited experience in organizing written texts. As Li and Hafner (2021) note, such support enables learners to focus their cognitive resources on content development and linguistic choices, rather than struggling with unfamiliar textual formats.

However, genre awareness should not be treated as a formulaic skill. Teachers must guide students to critically examine AI-generated texts, recognizing that genre conventions are context-dependent and shaped by audience, purpose, and cultural norms. Rather than passively adopting AI suggestions, learners should be encouraged to evaluate them against authentic models and their own communicative intentions.

Vocabulary and language development

Vocabulary acquisition and syntactic development remain core aspects of language learning in TESOL classrooms. GenAI tools provide learners with opportunities to expand their lexical range and experiment with new syntactic forms in meaningful contexts. By offering alternative word choices, paraphrasing suggestions, and context-sensitive synonyms, AI tools promote greater lexical variety and fluency in writing.

For instance, a learner working on a multimodal project may input a draft script or narrative and ask the AI to suggest more precise verbs, academic collocations, or sentence transitions. Such interactions promote lexical awareness and support the development of vocabulary depth and appropriacy. Ziegler (2023) highlights that these AI interactions, when combined with teacher feedback, can enhance both lexical diversity and learner confidence in word choice.

Moreover, GenAI tools can provide explanations or contextual examples for unfamiliar vocabulary, helping students acquire new terms through situated learning. Learners might ask the AI to define or demonstrate the use of a word in different genres, promoting deeper semantic understanding. This aligns with task-based language teaching approaches, where vocabulary is learned through purposeful communication and contextual relevance.

Syntactically, AI tools can support development by identifying run-on sentences, suggesting complex sentence structures, or transforming simple sentences into more varied constructions. Lin and Warschauer (2023) argue that AI tools that highlight syntactic alternatives and explain their pragmatic value can enhance learners’ syntactic range while preserving their personal voice. However, as with vocabulary, the role of the teacher remains crucial in guiding learners to apply AI suggestions selectively and critically.

To avoid over-dependence, educators should encourage students to engage in reflective language practices. Activities such as maintaining revision logs, comparing drafts, or discussing the rationale for vocabulary changes can foster metalinguistic awareness and promote long-term language development.

Multilingual and multimodal alignment

In multilingual TESOL contexts, learners may compose multimodal texts that draw on more than one language, especially during planning or idea generation phases. GenAI tools can support this multilingual dimension by translating content, offering comparative linguistic examples, or helping learners shift between languages in a controlled manner.

For example, a student may begin drafting a video narration in their first language (L1), then use GenAI to translate or paraphrase the content into English (L2), maintaining meaning while adapting it to academic or genre-specific conventions. This cross-linguistic support can reinforce language transfer and scaffold the movement from ideas to finished products. Li and Hafner (2021) emphasize that multilingual learners benefit from tools that validate their full linguistic repertoire while helping them develop proficiency in English.

Multimodality adds another layer of complexity. Learners must ensure that the various elements in their compositions—text, images, audio, video—work together to communicate a coherent message. GenAI tools can assist in evaluating this alignment by checking whether captions accurately describe visual content, whether spoken text complements background visuals, or whether written narratives flow logically with accompanying design elements.

For instance, a learner preparing a digital story can use GenAI to summarize the story’s main message, then compare that summary to the visual sequencing of slides or images. This process helps identify mismatches or redundancies between linguistic and visual modes. GenAI tools can also help optimize text density in slides or reduce redundancy between narration and on-screen text, supporting a better user experience and comprehension.

Educators can design guided tasks that prompt learners to use AI to evaluate multimodal cohesion. For example, students might use an AI-generated checklist to review their projects, asking questions such as: “Does each visual element support the main idea?” or “Are the transitions between images and text smooth and logical?” This not only improves the final product but also enhances learners’ understanding of how language and other modes interact in meaning-making.

Nevertheless, AI cannot fully replace teacher insight in addressing cultural nuance, rhetorical variation, and learner intention. Teachers must scaffold students in critically evaluating AI outputs, ensuring that their multimodal and multilingual choices reflect not just functional alignment but also purposeful and meaningful communication.

Conclusion to Section

Generative AI tools offer powerful scaffolding opportunities for TESOL learners engaged in multimodal composing. By modeling genre structures, expanding vocabulary, and supporting multilingual and multimodal integration, these tools can help learners navigate complex communicative demands while developing confidence and competence in English.

However, effective use requires pedagogical oversight. Teachers must guide learners in critically engaging with GenAI suggestions, promoting awareness of context, audience, and purpose. Through structured activities, reflection, and human guidance, GenAI can become a valuable partner in language and content development, enhancing learner agency while supporting long-term growth.

7. Addressing Authenticity and Assessment Validity Concerns

Risks of AI-generated content in assessment

As Generative AI tools become increasingly integrated into TESOL classrooms, educators are facing new challenges regarding the authenticity of student work and the validity of assessments that may be influenced by AI assistance. While these tools offer support for language development and multimodal composition, they can also generate entire texts, rewrite drafts, and provide extensive corrections—all of which blur the line between learner-authored output and AI-assisted production.

One of the key concerns is that students may rely too heavily on GenAI tools to complete assignments, particularly those that involve complex language or unfamiliar genres. This can lead to final submissions that reflect AI capabilities more than the learner’s actual proficiency. In such cases, assessment outcomes may not provide an accurate picture of student progress or ability (Dawson, 2023). For educators, this raises serious questions about the validity of grades, as well as the reliability of feedback based on work that may not be entirely the student’s own.

Kohnke et al. (2023) also note the risk of unintended plagiarism. AI-generated content can sometimes reproduce familiar phrasing or unoriginal ideas without proper attribution, especially when learners do not fully understand how to interpret or revise the content they receive from GenAI platforms. Without clear guidance, students might unknowingly submit work that is not fully their own, undermining the academic integrity of writing tasks.

Additionally, the lack of transparency in GenAI processes contributes to the problem. Most tools do not disclose the sources or reasoning behind their outputs, making it difficult for teachers to evaluate how much influence AI has had on the student’s final product. In multimodal contexts, where learners combine multiple elements such as text, image, and audio, identifying the boundaries between learner authorship and AI-generated support becomes even more complex.

Strategies for promoting learner-authored work

To address these challenges, educators can implement a range of strategies that promote authenticity and support valid assessment practices. One of the most effective is emphasizing process-based writing. Rather than evaluating only final products, teachers can assess writing development over time through drafts, outlines, annotations, and revision logs. This approach highlights the learner’s individual journey and allows instructors to observe language growth in context (Sowell, 2022).

Requiring students to submit multiple drafts, along with reflections on how and why they revised their work, encourages ownership and makes visible the role that AI tools may have played in the writing process. Reflection prompts can ask learners to identify specific feedback received from AI, describe how they responded to it, and explain what they learned from the interaction. This type of metacognitive engagement not only reinforces learning but also fosters academic honesty.

Another helpful practice is integrating in-class writing tasks that are not supported by AI. When students compose shorter texts under timed or supervised conditions, teachers can more accurately gauge language proficiency. These low-stakes activities provide useful benchmarks for comparing independent performance with AI-assisted work submitted for formal assessment. Dawson (2023) suggests that such triangulation of assessment data helps validate learning outcomes while minimizing the influence of AI on core performance indicators.

Clear and transparent guidelines regarding the acceptable use of GenAI tools are also essential. Teachers should explicitly communicate when and how AI support may be used—for example, for brainstorming, outlining, or receiving feedback—but clarify that final writing should reflect the learner’s own ideas and voice. Involving students in the development of these policies can increase their understanding and commitment to responsible use.

Additionally, the inclusion of oral presentations, peer reviews, and one-on-one writing conferences provides further insight into learner engagement and authorship. These supplementary activities create opportunities for students to demonstrate understanding, discuss their writing choices, and clarify how they used AI tools during the composing process.

Teacher agency and learner trust

Teacher agency plays a central role in managing the ethical and pedagogical dimensions of AI-assisted learning. Educators must exercise professional judgment in interpreting AI-influenced texts and remain actively involved in assessing not just outcomes but also the process by which they are achieved. This involves being attuned to individual learner profiles, setting clear expectations, and creating a classroom culture that values learning over perfection.

Firth (2023) argues that teachers must be empowered to adapt institutional policies and design instructional frameworks that reflect the realities of GenAI use. These frameworks should support flexible assessment models, allow for student self-assessment, and prioritize ongoing dialogue between teacher and learner. By maintaining visibility into the learning process, teachers can ensure that their feedback and evaluations are both fair and personalized.

Building learner trust is also crucial. Students are more likely to use AI tools responsibly when they understand the rationale behind classroom policies and feel supported rather than policed. Transparent communication about the affordances and limitations of GenAI fosters a respectful learning environment where students are encouraged to develop skills rather than simply submit polished products.

Moreover, trust is reinforced when students perceive assessment as an opportunity for feedback and improvement rather than as a judgment of ability. By emphasizing formative feedback, providing options for revision, and acknowledging the challenges of writing in a second language, teachers help learners feel secure in experimenting, making mistakes, and growing through guided support.

Ultimately, safeguarding authenticity in AI-assisted learning requires a balanced approach. GenAI tools should be seen as one resource among many—useful for prompting reflection, enhancing language accuracy, and modeling structure—but not as substitutes for learner effort or teacher expertise. Through thoughtful scaffolding, process-based tasks, and open communication, TESOL educators can ensure that writing instruction remains focused on meaningful learning and that assessment continues to reflect genuine student achievement.

Best Practices and Recommendations for TESOL Educators

Integrating GenAI ethically and effectively

As Generative AI becomes more accessible in educational settings, TESOL educators are tasked with determining how to use it ethically and effectively. Successful integration involves a combination of sound pedagogical planning, awareness of institutional expectations, and sensitivity to learner needs. GenAI tools offer meaningful support for digital multimodal composing (DMC), but their implementation must be grounded in a framework that promotes transparency, respect for student work, and constructive engagement with technology.

An ethical approach begins with clearly defining the role of GenAI within the classroom. Educators should establish guidelines that clarify when and how AI tools may be used. For instance, students might be permitted to use GenAI for brainstorming or organizing ideas but required to draft and revise text independently. Clear expectations help prevent misuse and reinforce the value of learner agency in the writing process (TESOL International Association, 2023).

Educators should also critically evaluate the tools themselves. Not all GenAI platforms are designed with educational use in mind, and some may pose risks related to data privacy, content accuracy, or user accessibility. Teachers should prioritize tools that are transparent in their operations and offer safeguards for learner data. Institutions can support this effort by vetting recommended platforms and providing professional guidance on their responsible use.

Effective integration also depends on alignment with language learning objectives. GenAI should serve to enhance, not replace, essential pedagogical practices such as scaffolding, formative feedback, and collaborative learning. Selwyn (2022) cautions that AI must be seen as a complement to, rather than a substitute for, the professional expertise of educators. When implemented thoughtfully, GenAI can help streamline instruction, increase access to feedback, and support the writing process in both linguistic and multimodal dimensions.

Sample tasks and instructional workflows

To support effective classroom use of GenAI, educators can incorporate specific task designs that emphasize process, creativity, and critical thinking. Below are several practical examples of how GenAI can be integrated into multimodal writing instruction:

1. AI-assisted brainstorming and outlining
Learners can begin a digital storytelling project by prompting an AI tool to generate topic ideas or a basic structure for their narrative. They can then modify the suggestions to suit their goals and audience. Teachers can scaffold this activity by providing genre-specific input and encouraging learners to justify their choices.

2. Language-focused revision workshops
Students can input their drafts into a GenAI tool and receive feedback on sentence structure, vocabulary, and tone. In pairs or small groups, they compare AI suggestions with peer feedback and decide which changes to implement. This promotes metalinguistic awareness and collaborative reflection.

3. Evaluating multimodal coherence
After assembling a digital composition, students use AI-generated questions or checklists to assess the alignment between visual and textual elements. For example, does the image enhance or distract from the written message? Does the audio narration complement the visual sequence? These questions help learners fine-tune their projects.

4. Reflective writing on AI use
Learners write brief reflections on how they used GenAI during a project. What suggestions did they accept or reject? How did the tool influence their final product? These responses provide insight into the learner’s writing process and help teachers assess authenticity and engagement.

Such task designs encourage active learning while embedding AI use in a pedagogically sound workflow. Teachers remain central in facilitating these activities, offering instruction, clarification, and support as students interact with both technology and content.

Professional development and policy alignment

Sustainable integration of GenAI into TESOL instruction requires continued investment in teacher training and policy alignment. As new tools emerge and existing platforms evolve, educators must stay informed about their capabilities, limitations, and implications for teaching and assessment.

Professional development initiatives should focus on both technical and pedagogical dimensions of AI use. Heift and Vyatkina (2022) recommend training that helps teachers understand how AI tools generate content, what types of feedback they can offer, and how to critically evaluate their output. Such knowledge enables teachers to make informed decisions and to model responsible use for students.

Collaborative learning communities also play a vital role. Teachers benefit from sharing experiences, comparing classroom strategies, and discussing ethical questions related to AI in education. Institutional support for such collaboration can take the form of workshops, online forums, or mentorship programs.

In addition, policy guidance from educational institutions is essential to ensure consistent and fair use of GenAI tools across programs. Clear policies should address acceptable use, data privacy, plagiarism prevention, and support for learners with varying levels of digital literacy. Ellis (2020) emphasizes that these policies should be developed in dialogue with educators to ensure that they are realistic, context-sensitive, and supportive of instructional goals.

Finally, adaptability and reflective practice remain key. As the landscape of GenAI continues to shift, TESOL educators must be prepared to adjust their strategies, reevaluate their tools, and reflect on the impact of AI use on learning outcomes. Periodic review of instructional practices, informed by student feedback and ongoing research, ensures that GenAI integration remains relevant and effective.

Conclusion

The integration of Generative AI into TESOL writing instruction marks a significant shift in how language learning and composition are approached, particularly within the expanding domain of digital multimodal composing (DMC). As explored throughout this article, GenAI tools offer a wide range of pedagogical affordances—from supporting genre awareness and scaffolding vocabulary to enhancing formative feedback and assisting with rubric-based evaluation. These capabilities, when thoughtfully applied, can help educators better support learners in developing both linguistic competence and multimodal literacy.

However, the benefits of GenAI must be balanced with careful consideration of instructional integrity and learner development. While automation can improve efficiency and access to support, the role of the teacher remains central in shaping meaningful learning experiences. Teacher mediation ensures that AI-generated input is aligned with educational goals, responsive to student needs, and grounded in pedagogical reasoning. It is through this informed facilitation that GenAI becomes a useful tool rather than a substitute for thoughtful instruction.

Moreover, maintaining authenticity and assessment validity requires proactive strategies. Process-based writing, reflection, and transparent discussions about AI use help ensure that learner-authored work remains at the center of evaluation. These practices not only promote trust between teacher and student but also foster greater learner agency. When students are guided to engage critically and ethically with GenAI tools, they develop skills that go beyond task completion—skills that support sustained growth as communicators in academic and professional settings.

The successful integration of GenAI also depends on institutional support, clear policy development, and ongoing professional learning. As tools continue to evolve, educators must remain adaptive, critically engaged, and committed to reflective practice. The dynamic relationship between technology and pedagogy means that TESOL professionals must continuously assess the impact of innovation on their classrooms, maintaining a steady focus on respectful, learner-centered teaching.

In sum, Generative AI holds considerable promise for enhancing multimodal writing instruction in TESOL contexts. Yet its effectiveness depends not on the technology itself, but on the decisions educators make in how they implement it. By aligning AI use with core pedagogical values and maintaining a clear commitment to learner development, TESOL educators can harness GenAI as a meaningful resource for supporting writing, fostering creativity, and promoting thoughtful engagement in multilingual, multimodal communication.

References

Ajjawi, R., & Boud, D. (2017). Exploring feedback dialogue: A model for effective feedback in higher education. Assessment & Evaluation in Higher Education, 42(4), 552–564. https://doi.org/10.1080/02602938.2016.1236361

Coniam, D. (2021). Using artificial intelligence to support writing assessment: Evaluating benefits and challenges. Language Learning & Technology, 25(3), 1–18. https://doi.org/10125/73488

Dawson, P. (2023). Generative AI and assessment legitimacy: Risks and responses. Assessment & Evaluation in Higher Education. Advance online publication. https://doi.org/10.1080/02602938.2023.2222583

Elgort, I. (2022). Artificial intelligence in second language learning: Prospects and pitfalls. Language Learning & Technology, 26(1), 1–7. https://doi.org/10125/73441

Ellis, R. (2020). Reflections on task-based language teaching. Multilingual Matters.

Firth, G. (2023). Navigating AI in the writing classroom: Professional agency and student engagement. English Teaching & Learning, 47(3), 203–219. https://doi.org/10.1007/s42321-023-00123-y

Godwin-Jones, R. (2023). Artificial intelligence and language learning: The new frontier. Language Learning & Technology, 27(1), 1–12. https://doi.org/10.55626/llt.2023.27.1.1

Hafner, C. A. (2021). Promoting digital multimodal composing in English language education: A review of the literature. Language Learning & Technology, 25(2), 24–44. https://doi.org/10125/73449

Hafner, C. A., & Ho, W. Y. J. (2020). Teaching and assessing digital multimodal composing: Towards a metalanguage and pedagogical framework. TESOL Quarterly, 54(4), 988–1011. https://doi.org/10.1002/tesq.579

Heift, T., & Vyatkina, N. (2022). Intelligent CALL and learner modeling: Supporting teacher agency in the age of AI. Language Learning & Technology, 26(2), 1–17. https://doi.org/10125/73450

Kohnke, L., & Moorhouse, B. L. (2023). Exploring the role of generative AI tools in English language teaching and learning. RELC Journal. Advance online publication. https://doi.org/10.1177/00336882231202975

Kohnke, L., Zou, D., & Xie, H. (2023). Ethical implications of generative AI in second language education. Computer Assisted Language Learning. Advance online publication. https://doi.org/10.1080/09588221.2023.2229950

Lee, J. S. (2022). Enhancing L2 writing feedback with artificial intelligence: Student engagement and learning outcomes. Language Teaching Research, 26(5), 785–805. https://doi.org/10.1177/1362168820976790

Li, J., & Hafner, C. A. (2021). Multimodal literacy engagement in the English language classroom: Affordances of artificial intelligence. System, 103, 102660. https://doi.org/10.1016/j.system.2021.102660

Lim, C. P. (2022). Leveraging artificial intelligence in assessment for learning: Conceptualizing human–AI interaction in language education. Computer Assisted Language Learning, 35(5–6), 856–873. https://doi.org/10.1080/09588221.2020.1868530

Lin, C.-H., & Warschauer, M. (2023). Language learning with AI writing tools: Affordances, limitations, and pedagogical considerations. Language Teaching Research. Advance online publication. https://doi.org/10.1177/13621688231174619

New London Group. (1996). A pedagogy of multiliteracies: Designing social futures. Harvard Educational Review, 66(1), 60–92. https://doi.org/10.17763/haer.66.1.17370n67v22j160u

Ortega, L. (2017). New CALL-SLA research interfaces for the twenty-first century: Toward equitable multilingualism. CALICO Journal, 34(3), 283–316. https://doi.org/10.1558/cj.27624

Ranalli, J. (2021). Automated written corrective feedback: How might it support learner development? Language Learning & Technology, 25(1), 1–23. https://doi.org/10125/44714

Reinhardt, J. (2020). Social media in second and foreign language teaching and learning: Blogs, wikis, and social networking. Language Learning & Technology, 24(1), 1–25. https://doi.org/10125/44718

Schmidt, R. (1990). The role of consciousness in second language learning. Applied Linguistics, 11(2), 129–158. https://doi.org/10.1093/applin/11.2.129

Selwyn, N. (2022). Should robots replace teachers? AI and the future of education. Polity Press.

Smith, B. E., & Craig, H. K. (2020). Multimodal composing and digital storytelling in TESOL: Towards a pedagogy of meaning-making. TESOL Journal, 11(1), e00454. https://doi.org/10.1002/tesj.454

Sowell, J. (2022). Process writing in EFL contexts: Fostering student voice and reducing plagiarism. TESOL Journal, 13(4), e00603. https://doi.org/10.1002/tesj.603

TESOL International Association. (2023). Position statement on the use of generative AI in English language teaching. https://www.tesol.org/docs/default-source/advocacy/position-statement-generative-ai.pdf

Wang, Y., Luo, J., & Sun, Y. (2023). Exploring AI-assisted multimodal feedback in EFL writing instruction. Computer Assisted Language Learning. Advance online publication. https://doi.org/10.1080/09588221.2023.2208121

Warschauer, M., & Liaw, M. L. (2023). Chatbots and AI in TESOL: Opportunities and limitations. TESOL Journal, 14(2), e00729. https://doi.org/10.1002/tesj.729

Ziegler, N. (2023). Harnessing AI for lexical development in instructed SLA: Risks and opportunities. Annual Review of Applied Linguistics, 43, 63–80. https://doi.org/10.1017/S0267190523000047