In the age of generative AI, writing classrooms and online platforms are facing a subtle but powerful phenomenon: linguistic flattening. I have been noticing it more and more. The way writers, consciously or unconsciously, start to lean on the same phrases, sentence structures, and tonal conventions is concerning. Undergraduate students, especially those for whom English is an additional language, are particularly vulnerable. AI offers quick fixes to grammar correction, stylistic polish, surface-level coherence, but it cannot replace the kind of cognitive work that essentially develops writing skills. As Lynda Gratton (2024) points out, AI may accelerate learning, but it cannot replace the development that shapes expertise, judgment, and identity. Operating with its own assumptions about “good writing,” AI can also encourage multilingual students to suppress their linguistic voices in favor of standardized American English, which is widely rewarded and privileged, because “histories of linguistic ideologies and dominant power” have “oppressed, mitigated, and erased marginalized communities” and their linguistic repertoires (Byrd, 2023, p.136). In doing so, many students unknowingly flatten the embodied, translingual practices that make their work culturally grounded, dynamic, and intellectually rich.
This is not just about written essays. In digital spaces, where writing increasingly mixes with design, sound, video, and interactivity, students’ digital and multimodal literacies are at risk. They often feel pressure to adopt familiar, conventional and predictable design choices, just to make their projects “legible” or “acceptable” to their audience by sidelining experimentation, culturally specific visuals, or unconventional storytelling. As Fernandes and McIntyre (2025) note, “linguistic and algorithmic oppressions are inextricably linked to one another,” standardization in both textual and digital worlds are now flattening students’ voices. While students may know they “have the right” to their own language, in practice, they often experience that points are awarded for grammatical correctness, surface-level coherence, and conventional design, sending a clear message about what counts as success. Hence, the pressure to sound and look “perfect” often leads students to over-rely on AI because writing support tools such as ChatGPT are“powered by language models that reflect social biases” and tend to homogenize writing toward Western styles, diminishing cultural nuances (Agarwal et. al. 2025).
In my class, I have noticed that students, particularly multilingual students, who may feel that their own patterns of language or culturally grounded design are “incorrect” or “unacceptable” or simply risky, tend to rely more on AI-generated content. To address this concern of linguistic and visual flattening in an algorithmic world, I integrated translingual pedagogy into multimodal literacy. While translingual pedagogy emphasizes valuing students’ full range of linguistic repertoire such as their dialects, and culturally specific modes of expression rather than enforcing a single “standard” English, multimodality honors students’ diverse learning and expression styles. This combination centers students’ own voices and encourages experimentation with text, visuals, and multimodal storytelling in their own style. Students negotiate meaning critically, make intentional multimodal and linguistic choices, and resist pressures to standardize their work according to AI or institutional norms. In doing so, translingual multimodal pedagogy promotes social justice by validating marginalized ways of knowing and communicating, and contributes to human-centered praxis that foregrounds students’ creativity and cultural specificity over AI-centered standardization.
Based on this translingual multimodal praxis, in my first year writing course, I teach the final paper: A Multimodal Remix project to encourage students to go beyond traditional writing practices. This assignment is liberating for many students who find designing their thoughts more exciting than textually writing about it. This assignment also helps students find feasible ways to honor students’ unique expression styles. For this assignment, students take an already completed assignment such as a personal narrative, rhetorical analysis, or synthesis paper, and transform it into a multimodal digital project such as an infographic, or a podcast, or a multimodal blog for a website or a short vlog.
I clearly mention that students are encouraged to practice adapting academic work for public, non-academic global audiences and develop skills in visual communication and multimodal composition. The word “global” brings the idea of “Global Englishes” back to our discussion. Hence, the importance of cultural representation also becomes evident. This is a public-facing, conversational, open-access genre, aimed at a general audience who may use English as an additional language. So, students are encouraged to use language that may or may not follow standard American English. I further encourage students to make intentional design choices to keep their message comprehensible and accessible to their target audience.
Pedagogically, this assignment allows me to discuss how multimodal rhetorical elements such as hyperlinks, images, colors, layouts, and content shape arguments and influence audience perception. It also opens conversations about ethical persuasion, which is increasingly important in a digital environment that is saturated with AI-generated misinformation and disinformation. For example, one of my students created a social media post celebrating their current or future professional achievement. The student generated an image of a well-known landmark using ChatGPT to avoid copyright issues, but the AI misrepresented both the architecture and the name of the building. What was meant to be a celebratory post unintentionally became a piece of misinformation.
This moment became a teaching opportunity for me to discuss how AI can hallucinate not only textual facts but also visual representations of reality. More importantly, it helped students recognize that digital writing often circulates beyond local classrooms. Posts on platforms like LinkedIn or other social media can reach global audiences who may rely on these visual and textual cues as credible information. That means digital writers are held responsible for creating and circulating publicly available information and knowledge. To enhance accessibility of the digital contents, I encourage students to use Global Englishes for multilingual audiences, provide alt-text for images if needed, ensure readable color contrast, and avoid culturally specific references that may not translate across contexts or at least provide related contexts to understand cross-cultural communication.
Another interesting remix project is audio content, such as podcasts. I have noticed that students often become more mindful of their global audience, who may come from anywhere in the world, using English as a first or additional language, when they use verbal language. They slow down their speech, articulate more carefully, and pay closer attention to clarity. In case of not having clear understanding of cross-cultural contexts, I encourage them to experiment with AI as a tool for exploring intercultural competence, which is the ability to recognize cultural differences, understand how meaning shifts across contexts, and communicate respectfully with diverse audiences, but never to write their scripts. I ask them to consider me as their global audience. I provide brief context about my cultural and linguistic background and encourage them to use AI to explore additional cultural nuances. Since I am accessible to my students, they can also check with me directly, and I can clarify their questions and concerns. This approach positions AI as a starting point for reflection rather than a cultural authority, since generative AI reflects dominant linguistic and cultural norms when prompted in specific languages and can potentially perpetuate cultural flattening.
This becomes an important teaching opportunity for me to discuss how AI operates in cross-cultural and intercultural communication, what it produces and how closely it aligns with real-world representations, whose perspectives are included or omitted, and how visual or textual choices may affect accessibility and inclusivity for a global audience. When I position myself as their “global” audience, it becomes easier for students to cross-check their work with me. However, what happens when their audience extends beyond the classroom and beyond cultures familiar to us? This question underscores the importance of human communication, interaction, and cross-cultural collaboration in sustaining cultural and linguistic nuances. Finally, when students submit their multimodal projects, they include a reflective author’s note that makes their thinking, writing, revising, and AI integration process visible. This note helps them develop metacognitive awareness and intentional, culturally grounded composition practices.
While Multimodal Remix assignments are becoming common in writing classrooms, the ways instructors intervene are crucial for fostering linguistic justice and countering AI-induced flattening. Based on my experience, there are four key strategies that I can recommend to fellow instructors:
- Design assignments that foster critical AI engagement: Teachers can encourage students to reflect on both textual and visual AI outputs, considering audience, context, and ethical persuasion.
- Provide guidance that honors effort and supports translingual experimentation: We can help students expand their linguistic and multimodal repertoires rather than conform to standardized norms.
- Intervene thoughtfully during the writing process: It is incredibly crucial to prompt students to consider AI’s limitations and biases, encouraging reflection on when and how to use AI responsibly and when not to use it at all.
- Cultivate an inclusive classroom environment: Lastly, it is important to center student agency, celebrate diverse voices, and normalize experimentation across modes and languages. When students find comfort in the classroom, they find confidence in their voices and expressions.
By integrating these practices, instructors can empower students to use AI responsibly while maintaining the richness, cultural specificity, and global resonance of their work.
References:
Agarwal, D., Naaman, M., & Vashistha, A. (2025). AI suggestions homogenize writing toward Western styles and diminish cultural nuances. In Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI ’25) (pp. 1-21). ACM. https://doi.org/10.1145/3706598.3713564
Byrd, A. (2023). Truth-Telling: Critical Inquiries on LLMs and the Corpus Texts That Train Them. Composition Studies, 51(1), pp. 135-142. https://compstudiesjournal.com/wp-content/uploads/2023/06/byrd.pdf
Fernandes, M. and McIntyre, M. (2025). Recoveries and Reconsiderations: Linguistic Justice and Storying Resistance to Generative AI. Peitho, 27(2). DOI: 10.37514/PEI-J.2025.27.2.05
Gratton, L. (2025). AI is Changing How We Learn at Work. Harvard Business Review. https://hbr.org/2025/12/ai-is-changing-how-we-learn-at-work
Students’ Right to Their Own Language. (1974). CCC, XXV. https://cdn.ncte.org/nctefiles/groups/cccc/newsrtol.pdf?_gl=1*1r9rqy3*_gcl_au*NjM5MjI2ODI3LjE3NzA5Mjg1Mjk.
