I am more creative in Turkish, my mother tongue. My arguments are sharper, my reasoning more precise, my voice more my own. Yet the academic system demands that I perform in English, not as a pedagogical goal, but as an implicit condition for being taken seriously at all. Academia has never quite come to terms with what that costs. We now finally have a tool that can begin to correct it. Yet here we are, policing it.
English dominance in academia is not a natural condition. It is the product of history. Colonialism, trade, geopolitical power… A hundred years ago, French was the language of intellectual life. Before that, Latin. Before that, Arabic, which for centuries was the lingua franca of science and mathematics. The prestige language has always been the language of the prevailing power. It is not the most expressive one or not the most logical one at all but the most strategically advantageous one to master.
Here it is worth pausing on the irony for a second. Kant wrote in German. Ibn Khaldun in Arabic. Simone de Beauvoir in French. The anglophone academy translated, absorbed and canonised their work and then demanded that future contributors deliver in English. The ladder is pulled up behind those who climb it.
This is also familiar territory for Swedish academics. The Swedish language has retreated as a language of scholarship under publication pressure. Formulating complex arguments in a second language carries a cost that is measured not only in time, but in the nuance and depth that get lost along the way. A non-native English speaker spends cognitive energy on the medium of communication that a native speaker can spend on the ideas. That is not a neutral standard. It is a structural handicap dressed up as academic objectivity.
And now the conditions are changing fundamentally.
AI translation and AI-assisted writing make it possible for a researcher in Nairobi, in Uppsala, in Buenos Aires to formulate their ideas in the language where they think best and then communicate them with a precision and fluency that previously required either native-speaker talent or decades of language training. This is not cheating. This is equity.
Academia’s response so far has been defensive. There is anxiety about AI-generated text, about hallucinated references, about what counts as authentic scholarship. Yes, the concern is not baseless. Hallucinated references in a submitted manuscript are a real problem, no argument there. But that is a problem of oversight, not of AI assistance itself. A hallucinating AI is a technical problem under active resolution. AI assistance in text writing is not a problem at all. By focusing on the symptoms, we are missing the diagnosis.
”A hallucinating AI is a technical problem under active resolution. AI assistance in text writing is not a problem at all. By focusing on the symptoms, we are missing the diagnosis.”
English functioning as a structural gatekeeper is a justice problem that academia has actively failed to address. I can imagine an objection: a shared language enables shared understanding, peer review, international collaboration. That is true. But that argument supports having a common medium of communication. It does not support measuring intellectual capacity through linguistic fluency, and it does not support letting historical power structures determine whose voice reaches an audience.
One day, and that day is coming much faster than we think, it will be impossible to determine whether a text was written by a person in English, by a person in Swahili with AI assistance, or by a person in Swedish who revised an AI draft. What will we do then? We can either have prepared ourselves, or we can be left guarding a gate that has already been removed.
It is time to begin preparing now. That does not mean ignoring misuse. It means shifting focus: from policing how ideas are expressed, to evaluating the quality of the question asked, the rigour of the method, the honesty of the interpretation, and the significance of the contribution. It means investing in AI competence as a research skill, not treating it as a threat to academic integrity. It means taking seriously that AI is genuinely leveling an uneven playing field and that this is a good thing.
Mert Can Yilmaz
Research Engineer
