Digital Literacy and Language Teaching: Toward Converging Concepts and Agendas in the Age of Ubiquitous AI

1. Introduction

Digital literacy has become an important educational concern that has figured in many international frameworks and national curricula as a crucial component of what it means to be a citizen in today’s world. It has also been extensively discussed in relation to language teaching and learning, no doubt because many in the field see digital literacy as a matter of critical, multimodal, and creative communicative competence that language teachers are well positioned to foster (Weninger, 2020; Darvin, 2019). Yet despite a general agreement about the importance and relevance of digital literacy for language teaching and learning, there are also notable differences in how digital literacy is conceptualized exactly, and thus how it should be researched and fostered in the language classroom.

One common sentiment I have come across in my own work with language teachers is that media and digital literacy is something that we ‘add-on’ to language teaching. According to this view, language teachers’ primary responsibility is to ensure students gain proficiency in the language skills. Digital literacy is a ‘good to have’ but cannot jeopardize time away from the primary task of linguistic skills development. This phenomenon is not unique to digital literacy but has been observed with regard to other ‘literacies’ as well. For instance, writing about critical literacy in the primary grades in Australia, Comber (2001) lamented how critical literacy was relegated to being a special classroom activity and often seen as a privilege available only to higher ability students, instead of teachers embracing it as an orientation to all texts, which must pervade literacy teaching and learning. A similar sentiment concerning multiliteracies has been documented among Singaporean English teachers, where multiliteracies constitute a key focal area in the national curriculum. As Lim et al. (2022) found, English teachers tended to ‘essentialize’ multiliteracies as concrete skills during interviews and were often observed to focus primarily on language features of multimodal texts they introduced in their classrooms. This is a view of literacy as a ‘zero-sum game’ (Kern, 2015, p. 220); any addition of ‘new’ literacies takes away from functional aspects of reading and writing.

Treating literacies as somehow external to the ‘core businesses’ of language teaching is surprising, given the literacy turn which has sought to bring insights of the New London Group (1996) to bear on the scholarship and teaching of modern foreign languages. Warner and Dupuy (2018) saw this turn as emerging from previous efforts toward a more integrated approach to teaching language skills, as well as a recognition that the teaching of (foreign/second) languages must be couched within a critical understanding of culture and context from the very beginning (Kern, 2000). The last several years have also seen an expansion of research into digital literacy in second language (L2) education, documenting what it takes to think of language learning as literacy practice (e.g., Lomicka, 2020; Reinhardt, 2019; Solmaz & Reinhardt, 2024). Moreover, with generative artificial intelligence (AI) entering many language classrooms, researchers have also sought to conceptualize its role within a (digital) literacy approach to language learning (Darvin, 2025).

There is thus growing evidence that when conceptualized as social practice, digital literacy holds great potential as an overarching frame for language teaching and learning. Foundational to understanding digital literacy as social practice is the idea that any act of meaning-making (whether with humans or robots, in one’s first, second or third language) is shaped by the beliefs, identities and purposes that people bring to bear on social situations, including classrooms (Darvin & Hafner, 2022; Jones & Hafner, 2021; Weninger, 2023). Digital literacy thus entails critical communicative (linguistic, multimodal, cultural) competence to accomplish social goals online, orienting to existing codes and norms of digital spaces while remaining sensitive to situationally contingent aspects of meaning-making (Weninger, 2022). At the same time, human agency alone does not shape digital practices; technology has material properties and dimensions that profoundly affect humans’ digital engagement and meaning-making (Godwin-Jones, 2024; Leander & Burriss, 2020). This is a foundational feature of our postdigital world where human and non-human activity is entangled in spaces no longer clearly demarcated as virtual or physical (Jandrić et al., 2018). Viewing digital literacy as practice rather than something an individual ‘possesses’ is thus crucial because it highlights literacy as emerging from durable and contingent elements in technologically mediated communicative environments. It is in this sense that scholars often prefer the plural form ‘digital literacies’, to signal that being literate is not a fixed state one either has or lacks, but that it is rather a complex practice whose exact configuration shifts across diverse digital communicative contexts. ¹

Yet, despite much support from a growing body of scholarship that has connected L2 teaching with ideas of multiliteracies and digital literacies, classroom practice has lagged behind. Warner and Dupuy (2018) highlighted two main reasons for this: one having to do with tensions between teachers’ training and their deep-seated ideas about how to teach languages, which are often at odds with a literacies approach to understanding language use. They also point to current teaching materials, which frequently present language in a culturally neutral or non-transparent way, focusing on language structure within imagined contexts of use. But one could argue that, in fact, the barriers are much more systemic and far-reaching and have to do with disciplinary traditions that have guided not only the scholarship of L2 teaching but also the training of language teachers in many parts of the world. Efforts to ‘integrate’ critical, media, multi or digital literacies into L2 teaching are, in a way, misleading because they may obscure some fundamental differences in theoretical concepts and curricular emphases which preclude a straightforward amalgamation of the two traditions.

Notwithstanding these enduring tensions, recent developments in AI technologies have made the adoption of multiple/digital literacies in foreign language education the only viable path forward. Consumer-friendly chatbots powered by large language models excel at understanding (or at least ‘responding to’) and generating text in multiple modalities. Not only are ChatGPT, Gemini and the like proficient language users, but they can also take on personae, imitate styles, as well as act as language tutors – again, in every mode of meaning-making. In fact, available data indicate that many people are using generative AI chatbots to help them with language tasks; writing assistance is the most frequently requested task for ChatGPT (OpenAI, 2025). With such a powerful resource at our fingertips, some have asked whether AI will one day replace language teachers or eliminate the need to learn languages altogether. If we continue to approach foreign language education as mainly teaching students how to listen, speak, read and write in an L2, there is a serious risk that we will become obsolete when students can learn how to do that from a bot. To avoid this, we must find ways to reconceptualize foreign language teaching and learning in an AI-mediated communication landscape. But before we can articulate curricular and pedagogic alternatives to language learning as the acquisition of linguistic skills, we urgently need conceptual clarity and a measure of unity in the field so that engagement with AI is viewed as a matter of literacy rather than tool use.

The aim of this article is to contribute to this conceptual work of what it might take for language teachers to view AI and other digital technologies from the point of view of digital literacy. In so doing, I will narrow my focus on areas of scholarship and practice that have been central to the field of L2 teaching. Specifically, I will focus on four such areas: the language curriculum, the relationship between technology and language learning, the language learner, and the practice of language teaching. Within each of these areas, I will provide an overview of key issues and enduring notions, drawing on a broad set of literature that traverses L2 language learning and literacy scholarship. For each area, I also make a proposal for what we might need to think about if our goal is to realign language teaching and AI with an orientation to digital literacy. Throughout the article, I will be centrally concerned with the question: How should we rethink key notions within language teaching to account for AI within a digital literacies perspective? In order to answer this question, it is necessary to revisit concepts in a historical context so that barriers to conceptual alignment can be identified, which is what the article aims to do.

2. Rethinking the Language Curriculum

One obvious point of departure is to examine what constitutes the substance or subject matter of language teaching. This is a question of curriculum: what body of knowledge, what skills or competencies are to be understood as comprising a language curriculum, and how should they be organized? The answers to these questions are shaped by broader ideas about the purpose of language education, which impact curricular selections. As an epistemological, normative, and practical endeavor (Deng, 2015), curriculum is historically contingent, and this can be readily seen in the field of language teaching as well. For instance, well into the twentieth century, the purpose of foreign language teaching was for learners to appreciate the literature of the ‘target culture’ (Richards & Rogers, 2014). As such, foreign language curricula prioritized reading, writing and grammar, with a heavy emphasis on accuracy, and using students’ native language as the medium of instruction. This notion of language development as a gateway to ‘good’ literature was also observable in the emphasis on reading in first-language teaching contexts at the time, such as in the United States (Applebee, 1974). In the last fifty years, language education has certainly moved away from the aim of developing proficiency for the purpose of reading the literary canon.

3. Rethinking the Relationship between Technology and Language Learning

As mentioned in the Introduction, terms like digital literacy, critical literacy or media literacy in the context of language teaching have been understood (by some scholars as well as practitioners) as something one adds to language learning. This reflects the view that, as language teachers, our primary responsibility is ensuring students are proficient in the four skills, after which we can concern ourselves with things like critical literacy. A similar idea persists when it comes to technology and language teaching, whereby digital technology is seen as a tool for language learning rather than a context for social-semiotic action. When technology is seen as a tool to make language learning more fun or more effective, then digital literacy will likely be defined as the appropriate and competent use of various digital technologies to facilitate the learning or acquisition of languages – whether as a language teacher or a student. The field of computer-assisted language learning (CALL) has historically concerned itself with precisely these questions – the externality of tech evident in the noun phrase structure of the field’s name. However, this kind of separation of technology and language/literacy as ‘aspects’ of context is increasingly untenable, especially with AI, as I have argued in the previous section. To be sure, there has been a critical turn within CALL over the past several years, with scholars increasingly critiquing the technological determinism and instrumentalism that seems to have defined the field (Anwaruddin, 2019). Similarly, critical voices have pushed back against the neutrality of technology in language learning, pointing to how neoliberal, capitalist narratives about technology’s redemptive powers for everyone silence deep inequalities around access and bias and the uneven pedagogical benefits of technology across social groups and contexts (Ahmed, 2022; Darvin, 2018; Hellmich, 2019). These are important developments for CALL and language teaching research.

But for language teaching as a classroom pedagogic practice, it is equally important to make the switch from viewing digital technology as an aid or tool to understanding how it shapes the very language/semiotic practices which it facilitates, and especially with regard to AI. Warschauer et al. (2023), for instance, examining AI from a sociocultural perspective, recognize its role in mediating the human activity of writing. As such, they argue, AI shapes the practice of writing as well as the cognitive processes involved. Yet the idea of tool integration remains strong. For instance, there is much research on the use of social media in the language classroom and its positive effects on aspects of language learning, such as writing (Barrot, 2021). But in much of this research, there is a tendency to treat digital media as an independent variable. While this may be a methodological necessity to quantitatively measure its impact on performance in conventional language skills, it reveals a conceptualization of technology as mostly autonomous from processes of language learning. This perpetuates an instrumentalist view of technology while overlooking the fundamental differences around semiotic practices of using social media in school as opposed to out-of-school settings, and the implications of these differences for language learning. A further connected problem is that other types of technologies, such as digital gaming, have not gained widespread legitimacy, in part because they are seen as external to processes of meaning-making. As Blume (2019) argued, despite the strong potential of digital games as environments for meaningful interaction, foreign language teachers’ reluctance is evident, owing to their attitudes about games, the role of schools as sites for serious work, and an imbalance in expertise, which would reverse the traditional epistemic hierarchy between students and teachers.

Blume’s arguments exemplify the prevalent view that examines and evaluates digital technologies from the perspective of their instructional potential, rather than how each presents unique contexts for developing diverse language and semiotic skills as part of a broader digital literacy practice. Starting from digital literacies allows us to focus on how being online is first and foremost about accomplishing social-communicative goals and negotiating social relationships while orienting to norms and values accrued within cultures of use (Thorne, 2003). Literacies are always situated, which means different digital environments, as (micro)-cultural practices, may call for very different kinds of literate identities, goals and semiotic practices. The language and multimodal competencies required to become or be seen as game-literate are very different from what it takes to become a popular TikToker. This is why seeing technology as an independent variable that can aid the classroom development of language in a largely context-free manner does not do justice to the multiple literacies involved in digitally mediated communicative practices (Reinhardt, 2024; Reinhardt & Thorne, 2019).

Seen in this light, interacting with LLM-powered AI chatbots then becomes another digital literacy practice – and not merely a technology to improve the four skills. Just as the practice of gaming is a process of identity development involving literacy practices specific to this context (e.g., play literacies, information literacies, system literacies; Reinhardt, 2024), so too should we conceive of AI-human interactional environments as new contexts for new kinds of identities and (digital) literacies. There are now studies emerging (see in this Special Issue) that document this contextual process, producing rich accounts of language and identity development that can take place even within classrooms through engaging with AI as a digital literacy practice. The key insight for language teaching here is that teachers need to design AI digital literacy practices as meaningful communicative contexts for the classroom, rather than thinking about their role as being to select the best tool to improve language proficiency.

4. Rethinking Notions of the Language Learner

An important question for language teaching in the age of AI is how we want to conceptualize the language learner as a literate being. This, of course, is closely connected to how we conceive of literacy and the process of language learning and meaning-making in digital spaces, the focus of the previous sections. Western education has long been premised on the Enlightenment ideal of the rational human being – the Cartesian subject who locates truth in the thinking mind, distinct from and superior to the sensory body. This is visible in the notion of authorship, and more generally in the idea that the individual student is the fons et origo of meaningful communicative action. Most language teaching approaches have taken this as a starting point, although communicative language teaching recognized the importance of context (notably, one’s interlocutor and the task) and conceptualized the learner as co-creator of meaning in task-focused activities in the classroom (Richards, 2006). But when it comes to assessment, evidence of language learning is still ascertained in many educational contexts by students’ individual communicative ability and output, underscoring literacy as as individual cognitive act and creating backwash effects for classroom instruction. Language learning as an individual cognitive process has endured despite increasing challenges over the past decades, prompted to a large extent by technological developments.

Although Bartes famously declared the death of the author as early as 1977 (even earlier in the original French), it was the arrival of the internet that offered a key moment of reckoning concerning our ideas about text and authorship. New literacies scholars in the early 2000s documented the ways in which the new technology was spawning new digital spaces and practices whose emerging culture of meaning-making departed from our book-world model of literacy. Lankshear and Knobel (2003) discussed authorship and agency as distributed, supported by the digital ethos of a participatory culture where sharing, collaboration and remix, rather than authority and individual literate acts, were becoming the norm, epitomized by literate practices around Wikipedia and online fanfiction. Similarly, the expanding semiotic complexity of new digital spaces made it more apparent that communicative action emerges through the shared cognition of people and artifacts operating in social environments. Language learning in such semiotic ecologies (Thorne et al., 2012) occurred less as a result of conscious human effort than the (mostly) unintended outcome of distributed cognitive processes across cultural-material contexts. But this is not unique to digital spaces; Pennycook’s (2017, p. 278) notion of semiotic assemblage places language ‘outside the head’ and reconceptualizes it as a ‘distributed effect of a range of interacting objects, people and places’ whether online or offline.

With social media companies increasingly deploying algorithmic feeds by the mid-2010s to curate personalized content for users, scholars began to take a more critical look at the role of technology in assembling literate activity online (Carrington, 2018; Jones, 2021). Sociomaterialism and affect studies have emerged as novel theoretical lenses to make sense of the role of non-human computational systems as ‘material agents’ (Ehret, 2024) in literacy practices. Continuing with earlier theoretical postulations about distributed cognition, agency in this body of work is seen not as residing in the literate individual but rather as a ‘relational achievement’ (Leander & Burris, 2020, p. 1265) among human and non-human agents within an algorithmic culture. Parallel to this, multiliteracies scholars have pushed for an embodied understanding of literacy as a fundamentally sensory experience (Mills, 2015; Mills et al., 2023), especially given how new, technologically mediated literacies (e.g., virtual reality) increasingly involve the body.

Then in 2022, ChatGPT arrived and, with it, perhaps an even greater challenge to our ideas of the homo loquens uniquely capable of speech and complex semiotic behavior. Unsurprisingly, agency has been a key concept around which much discussion in the field has taken place. If robots are capable of producing language (in all its modalities), what is the role of the human, both in the narrow sense of human-AI interaction as well as at a more philosophical level? Godwin-Jones’ (2024) argument to conceptualize agency in the age of AI as shared or distributed is very much aligned with the field’s ongoing orientation to ecological approaches and to sociomaterialism (as mentioned above), and validates the co-agency of AI due to its ability to effect change autonomously. That students also orient to such a fluid and fluctuating notion of agency has been attested by empirical work, for instance, in Jiang et al.’s (2024) narrative study, where students acknowledged and valued ChatGPT’s role as agentive and integral to their research writing processes. Their perspectives, according to the authors, reflect what Latour (2004) termed the ‘fair position’ in that while recognizing technology’s agentive role, students remained critical of its problematic aspects and vigilant about overreliance on the tool. At the same time, Zheng’s (2025) study with Chinese international students highlights how distributed agency is not an unproblematic notion. Students in his study reported a heightened sense of linguistic insecurity and language-related anxiety induced by ChatGPT’s native authority as an expert in standard academic English.

Within a digital literacies perspective, we could reframe these arguments as connected to notions of identity. New Literacy Studies has long recognized identity as central to literacy, given that literacy ‘involves people in participation, interaction, relationships, and contexts, all of which have implications for how people make sense of themselves and others, identify, and are identified’ (Moje & Luke, 2009, p. 416). As the quote already makes it clear, identity in practices of literacy is not seen as an expression of unfettered human agency or some essentialized self, but rather as an always only partially human or individual accomplishment. Here, I draw on Bucholtz and Hall’s (2005) theoretical exposition on identity within sociocultural linguistics, in which they proposed, twenty years ago, distributed agency as a way to recognize that social-semiotic action is always accomplished as an amalgam of human intention, habitual behavior, as well as ideological processes and material structures acting on the human agent. Semiotic-ecological and digital literacy approaches to L2 learning thus converge in their conceptualization of the language learner, viewing participation in communicative action through the complementary lenses of agency and identity, both understood as distributed across social, material, spatial, temporal and semiotic contexts.

When it comes to the role of the language learner, the real challenge is therefore not in theorizations of agency or identity; it is in reckoning with the implications of this insight for language teaching practice. Many educational systems across the world are premised on the idea of the individual student as the source of intelligence, which can be developed and measured, and on the basis of which educational credentials are given and career trajectories determined. The four-skills curriculum and viewing technology and literacy as independent variables often operate and perpetuate such systems, as the goal is to maximize language competence, rather than capacity for communicative and transformative action. Crucially, assessment in K-12 education in key subjects rests on this assumption and educational leadership lacks the political will or curricular structures to deviate from it. Yet if AI is to become an integral part of our personal and work lives, we need to switch our focus to developing (and assessing) how learners are able to recognize, negotiate, and leverage co-agents to accomplish contextually sensitive and contingent social-semiotic goals within a broad conception of digitally literate individuals. I believe that, as language teachers and researchers, our immediate task is to translate these theoretical ideas into curricular and pedagogic principles for how we might go about this in language educational practice.

5. Reaffirming Language Teaching as an Ideologically Implicated Practice

That language teaching is not an ideologically neutral activity is hardly a new insight, given decades of critical applied linguistics research and theorizing. The ideological nature of language education has been discussed in relation to many aspects of language teaching, including language textbooks (e.g., Dendrinos, 1992), the native speaker as ideal L2 model (Cook, 1999), the reification of standard languages (Tollefson, 2007), the legacy of colonialism in ELT scholarship and practice (Pennycook, 2007), how classroom discourse positions minoritized learners (Martin-Jones & Saxena, 1996) and many more. What is more, this reckoning with our field’s ideological non-neutrality has also energized theory and research of language teaching practices. The decolonial turn has questioned enduring normative assumptions, including what language is, who should teach it, and how it should be taught. Kumaravadivelu (2016), frustrated with the continued discrimination against non-native speaker English teachers, contends that the only decolonial option for them is to act as a subaltern collective within their own contexts. Canagarajah (2025) questions our persistent focus on proficiency in named languages and instead makes the case for teaching ‘semiotic repertoires’ so students can participate in diverse communicative activities. Translanguaging as classroom pedagogy offers a way to put some of these decolonial ideas into classroom practice: to treat the multilingual teacher as the norm rather than an aberration and treat language as semiotic practice rather than mastery of four skills.

AI confounds matters significantly, because its most popular iterations – LLM-powered applications – are experts at language. This raises a whole host of questions concerning the ideological basis of AI’s ‘language expertise’. For instance, what notion(s) of language and language speakers do AI chatbots orient to? In this regard, Crompton et al.’s (2024) review study identified that artificial intelligence technology’s most concerning negative impact within English language teaching is its tendency to operate based on standardizing and essentializing notions of languages. Likewise, Lau’s (2025) study of three major voice assistants’ language options found that they reinforced standard and national language ideologies. Unsurprisingly, studies increasingly point to similarly restrictive assumptions underlying the notions of language speakers and users embedded in AI systems. Payne et al. (2024) critically examined AI-assisted accent modification, arguing that they perpetuate deficit-oriented raciolinguistic ideologies by positioning linguistically minoritized speakers as inferior. Worryingly, Liang et al. (2023) found that AI-detection software shows significant native speaker bias as the systems they tested produced more false positives on essays written by non-native speakers. There is also growing evidence that learners’ experiences and views are impacted by the machines’ racialized and native-speaker ideals. Curran et al.’s (2025) analysis of reviews of a popular AI language learning app surfaced clear indications of users’ acceptance of native speaker ideologies promulgated by the app’s feedback and scoring system. Finally, Jeon et al. (2024) examined primary-age learners’ attitude toward English as long-term users of a language learning app and found that, despite clear positive outcomes, learners’ discourse reflected a strong orientation toward normative and essentialized understanding of language, which in turn constrained their potential for further development.

A related question concerns how the emergence of AI complicates the ideological nature of language teaching and learning. We could argue that treating AI technology as a tool for language learning, as explicated above, is itself ideological, if we define ideology as a dominant interpretive framework for making sense of the world and for regulating our behavior within it (van Dijk, 1998). It is ideological because it serves the interests of an established (sub)field within language learning by ensuring its continuation. Challenging the notion of technology as a tool or independent variable can be seen as a threat to the continued existence of this field. It is also ideological because such a view sustains the entanglement of the language teaching field with the profit-oriented educational technologies industry. Specifically, when viewed as external to the practice of language teaching and learning rather than actively shaping it, (AI) technology is offered as a ready-made solution to many of the ‘problems’ language teaching has faced. Such a fundamental belief in technology’s ability to improve human life and outcomes is an ideological discourse that has pervaded education (Selwyn, 2014), including language education. A series of examples from the mid-twentieth century onward, such as language laboratories, educational broadcasting, language packages on CD-ROMs, internet-based innovations (e.g., class chats), and, most recently, mobile language learning via apps illustrate that the ground-breaking impact of new technological solutions is rarely as promised. This is due, in part, to the ideological assumption shared by these innovations that technological efficiency translates directly into ease of learning, and that such ease will inevitably lead to linguistic development. When viewed along this historical trajectory, AI is therefore simply the latest arrival to the ideologically loaded scene of (language) educational technologies.

Teaching and learning languages, with or without technology, has always been an ideological endeavor, and digital literacies as a curricular and pedagogic approach is well suited to interrogate the many ideological foundations of our field and practice. Digital literacies’ emphasis on fostering critical dispositions toward AI/technology as non-neutral agents in meaning-making enables us to actively work against AI becoming a conduit for recolonizing language education (Dovchin, 2024). Such dispositions can be fostered in classroom practice by critically exploring with students the various ways in which AI technologies are ideological. For instance, how voice options in AI text-to-voice generating software (e.g., Elevenlabs) and AI voice systems more generally perpetuate racial, ethnic and gender stereotypes (Sindoni, 2024). We can draw on existing concepts, such as representation, to explore with students how AI image-generating platforms produce biased visual representations (Weninger & Xu, 2026) or how the materiality of generative LLM chatbots like ChatGPT or Claude significantly shape learners’ interactions and dispositions toward them (Darvin, 2025). All of these require that we conceptualize AI not as a tool but as a context for meaning-making – and a thoroughly social and ideological one at that. At the same time, we can focus on AI’s (co-)creative potential to disrupt current recolonizing tendencies documented by the above-reviewed studies. Bogachenko et al.’s (2024) study with people displaced due to the war in Ukraine illustrates how technology can be experienced as fostering confidence and a sense of freedom within otherwise traumatizing circumstances. In their study, participants’ engagement with Google Translate was associated with increased metalinguistic awareness, alongside critical dispositions as they recognized the technology’s limited awareness of linguistic diversity. We need more empirical studies that document digital literacy as a language classroom practice to explore the full experience of meaning-making with/in AI.

6. Conclusions

This article set out to reexamine core assumptions in L2 teaching in an age when AI is becoming ubiquitous both in our private and professional communicational landscapes. Specifically, I argued that digital literacies offer a compelling theoretical and practical frame to accommodate AI into our scholarly and professional practice. Touching on four areas – the language curriculum, the relationship between technology and language teaching, the language learner, and the ideological nature of language teaching – the article identified enduring scholarly conceptualizations and professional orientations which may hinder a shift toward digital literacies as a central framework for language teaching and learning. While the discussion remained primarily conceptual, I contend that discussing the impact of AI on language teaching this way is essential for surfacing theoretical and ideological fissures that keep us from converging toward a more pluralistic and critically social approach to language and literacy education.

To move forward, we must see AI not just as a disruption, as it is often framed, but also as an opportunity to revisit and rewrite enduring assumptions undergirding our scholarly research, thinking and professional practice. The fundamental question we are facing with AI is the same Kern (2015, p. 221) articulated more than ten years ago: do we ‘value and reflect the “multi” or . . . attempt to suppress it by emphasizing standardization and normativity in language use and meaning-making practices?’ For scholarship, valuing ‘the multi’ means continuing to develop concepts that embrace plurality and complexity in describing communication and language learning in AI-mediated contexts. An emphasis on literacy as process (Kalantzis & Cope, 2020), transmodal literacy, contextual awareness, distributed agency, and technologies as semiotic environments rather than tools are notable developments in the right direction. We need to put more effort into expanding our conceptual repertoire, not for the sake of creating ever-newer concepts, but to capture the layered meaning-making practices in digital times. As for professional practice, we need to press on with classroom-based empirical research to demonstrate the feasibility and desirability of digital literacies as an expansive approach to language teaching and learning. This is difficult given a global tendency toward reductive practices in formal language education triggered by the washback effect of standardized testing international competition has brought on (Weninger, 2019). But perhaps the ‘AI will disrupt everything’ narrative can be productively reframed as an opportunity for us to work with teachers, schools, districts and educational policy makers in advancing novel – and even bold – curricular and pedagogic approaches. If AI truly poses an existential challenge to established models of language learning and assessment, then this may be precisely the moment to foreground digital literacies as critical, multimodal, socially-grounded semiotic practices central to curriculum and pedagogy in L2 language teaching and learning in the age of AI.