Conclusions and recommendations for future work

With this review, we set out to ask whether two decades of research have yielded definitions and theories of culture that offer better frameworks analyzing contemporary digital (intercultural) communication, whether digital technologies necessarily import particular cultural values, and how the language and disciplinary scope of the field have shifted. Across the corpus, we do find some clearer, more usable accounts of culture–moving from essentialist, nation-bound models to constructivist and practice-based views attentive to superdiversity, mediation, and power.

With respect to theory, there is terminological progress but no final settlement. Terminology has modernized (from cyber to digital, from cyberspace to platformed worlds), and scholarly language now travels more fluently across media studies, journalism studies, sociolinguistics, HCI, and design–broadening both analytic tools and sites of inquiry. Increased attention has been given to multilingual and translingual practice, pragmatic work (stance-taking, mitigation, repair), and the role of platform norms in shaping interaction.

At the same time, the field still works with multiple, sometimes incompatible, meanings of culture, and operationalization often lags behind theory. Intercultural and transcultural frameworks co-exist, with both still sometimes treated as static binaries. Modern process-oriented theories of cultures–such as the transdifference approach that is vibrant in cultural and literature studies–are completely absent from the literature we located. Future research on culture and communication in digital worlds must continue to explore and embrace dynamic, processual approaches to the study of cultures and (trans-)culturalization.

Methods remain predominantly qualitative; comparatively little cumulative, longitudinal, or mixed methods research ties theoretical claims to actual comparable measures across languages, genres, and settings. Strengthening the field will require studies that make their cultural constructs explicit, align methods to those constructs, and report measures that permit cross-study synthesis.

Do digital technologies necessarily import particular cultural values? The evidence suggests a more qualified claim: technologies are not neutral, but importation is neither uniform nor unavoidable. Platform affordances, governance arrangements, data availability, and gatekeeping practices often channel participation toward dominant styles and centres of power; yet users also hybridize, localize, and resist. Our Fourth Age coda underscores how generative models intensify both tendencies: model defaults can normalize toward Western/educated registers, while targetted fine-tuning, co-design, and locally governed data practices can open room for plural “AI cultures.” The task ahead is to study–not assume–the conditions under which each trajectory prevails.

Accordingly, we recommend that future work proceed programmatically rather than piecemeal. Conceptually, researchers should state what they mean by culture (Inherited traits? Interactional practice? Media-shaped meaning-making?) and select designs that measure those commitments in use. Empirically, the field needs comparative, longitudinal, and mixed methods studies that connect interactional outcomes (e.g., participation patterns, disagreement management, register drift) with platform rules, model defaults, and local norms across languages and scripts. Infrastructurally, collaboration with communities–particularly beyond the Anglophone world–should extend from data governance to evaluation, so that claims about “effective” or “ethical” practice are tested against situated goals. Theoretically, we should continue to braid insights from sociolinguistics, media and journalism studies, STS, design justice, and education to explain how power and practice travel together. Taken together, these steps position the field to describe (and not just imagine) digital cultures in their plurality, and to shape technologies and learning environments that recognize, rather than overwrite, that plurality.