AI and the New Hegemony of Language: A Warning Against Digital Imperialism -By Olasunkanmi Arowolo

Artificial Intelligence is often presented as a neutral force, a tool that brings efficiency, knowledge, and connectivity, but the reality is far more complex. AI is never built in a vacuum. It carries within it the assumptions, biases, and cultural orientations of those who design and train it. Nowhere is this clearer than in how AI systems handle language. A troubling pattern has emerged in which American English is privileged as the default, while other forms of English are marginalised or erased.
This is not simply a matter of spelling preferences or stylistic conventions. It signals something more profound: the consolidation of American cultural power through digital means. What we are witnessing is a subtle but far-reaching form of digital imperialism, in which technology becomes a vehicle for linguistic dominance.

How the Centre Marginalises the Edges

Large language models (LLMs) are trained on vast datasets dominated by American content. This includes news outlets, academic texts, entertainment media, and online forums. As a result, the English that emerges in AI systems is overwhelmingly American in tone, spelling, idiom, and punctuation.

Other forms of English are treated as deviations. British English, once seen as the reference point for global communication, is often presented as a variant that requires adjustment. Nigerian English, Indian English, South African English, and many more are rendered virtually invisible, even though millions of people use these varieties in everyday life.

This erasure is not just technical. It is cultural. To suggest that only one form of English is correct is to silence other histories, literatures, and ways of expressing thought. AI, in its current form, often reproduces precisely that silencing.

From Colonial Textbooks to Digital Algorithms

This pattern is not new. During the colonial era, language was one of the most effective tools of control. Empires spread their influence not only through military force or economic power but through the imposition of linguistic norms. Textbooks, missionary schools, and broadcasting institutions set standards for how subjects were expected to read, write, and speak.

Today, the mechanisms have changed but the outcome looks familiar. Instead of textbooks, we have training datasets. Instead of colonial schoolmasters, we have algorithms. Instead of radio, we have autocorrect features and AI-generated text. Each of these reinforces the idea that American English is the natural centre of authority while all others are peripheral.

The effect is subtle, yet powerful. A student in Lagos may draft an essay in Nigerian English, only for an AI-powered tool to flag their spelling as incorrect. A journalist in London may ask a chatbot to generate copy in British English, only for the system to revert to American conventions after a few sentences. These small moments add up to something larger: the normalisation of American linguistic norms as the universal standard.

Language, Punctuation, and the Invisible Politics of Style

Even the smallest details of language carry cultural weight. Punctuation is one example. The frequent use of the em dash in American writing has now been absorbed by AI-generated content, often displacing other marks that are more common in British or international English. What looks like a stylistic preference actually shapes the rhythm of prose and the way ideas are structured.

The rise of the em dash may appear trivial, but it illustrates the broader issue. When one set of linguistic habits is universalised through technology, it creates subtle pressure to conform. Over time, these pressures reshape the boundaries of what is considered polished, professional, or authoritative writing. What is at stake is not just punctuation, but the authority to define language itself.

Why This Matters

Language is never neutral. It is deeply tied to identity, belonging, and power. The privileging of American English in AI systems has several consequences.

First, it impoverishes communication. Nigerian English, for example, is rich with proverbs, rhythms, and cultural references that convey meaning in ways American idioms cannot. Indian English carries its own cadence and history. British English embodies centuries of literary and political tradition. To reduce these forms to anomalies is to flatten the richness of global expression.

Second, it undermines epistemic justice. If AI is increasingly used to mediate education, journalism, business, and even diplomacy, then linguistic biases will shape which voices are taken seriously. A student or researcher who writes in a localised form of English risks being deemed less credible, not because their ideas lack value but because the system they are working through privileges American usage.

Third, it represents a subtle form of cultural control. When American English is treated as the natural default, it reinforces the impression that American perspectives are universal. This is precisely how hegemony operates: not through overt force, but by presenting one particular way of speaking and thinking as common sense.

Examples of Digital Imperialism in Action

Consider the experience of international students who rely on AI tools for academic support. A British or Nigerian student may prefer to write with British spelling, but AI grammar checkers often override their choices. Similarly, professionals in non-American countries find that workplace communication software defaults to American spellings, subtly disciplining them into conformity.

Another example is publishing. Many AI-assisted editorial tools push manuscripts into American house styles, regardless of the intended audience. For authors seeking to reflect local identity through language, this becomes a battle not just of style but of cultural authenticity.

In journalism, AI-driven summarisation tools often rely on American media sources. As a result, even when global events are summarised, the framing is more likely to reflect American perspectives. This reinforces not only linguistic dominance but also narrative dominance.

Towards a Plural Digital Future

There are constructive alternatives. Datasets can be broadened to include diverse varieties of English. Communities of speakers can be involved in shaping how their languages are represented in AI systems. Tools can be designed to lock in user preferences in ways that are stable and reliable, rather than reverting to a default.

Policymakers and regulators also have a role to play. Language diversity should be treated as an essential element of cultural sovereignty, not a minor setting buried in user preferences. Education systems can help students understand that their local varieties of English are not inferior, but part of a wider global tapestry.

Cultural leaders and writers can also resist by consciously maintaining their linguistic identities in digital spaces. Every time a Nigerian author insists on Nigerian usage, or a British journalist insists on British spelling, it is an act of cultural affirmation.

The Choice Before Us

The rise of AI is reshaping global communication. The question is not whether language will change, but who will control that change. The dominance of American English through AI is not a matter of convenience. It is a quiet advance of digital imperialism, carried not by armies or flags, but by algorithms and autocorrect.

The choice for the rest of the world is clear. Either accept this flattening of English into one dominant form or insist on a digital future that respects plurality and protects linguistic diversity. To accept the former is to concede cultural authority to a single centre. To choose the latter is to defend the richness of global English and the dignity of those who speak it.

The future of AI will be the future of language. What is at stake is nothing less than whose voices will continue to count in the digital age.

About the Author

Olasunkanmi Arowolo, PhD (University of Kent), teaches journalism and media at the Faculty of Communication and Media Studies, Lagos State University, Nigeria. His work focuses on journalism, digital governance, and media policy, and he is a dedicated advocate for quality education. He is a British Academy-endorsed Global Talent. He can be contacted at oa@olaarowolo.com or via X (Twitter): @olaarowolo.