Dispatch
When translation is instant, does language learning still matter
As AI enables instant translation, universities reassess the role and value of language learning.
By Gauri Kohli
Dispatch
When translation is instant, does language learning still matter
As AI enables instant translation, universities reassess the role and value of language learning.
By Gauri Kohli
"If AI can translate, do universities still need to teach languages?"
"AI is entering an already shifting landscape shaped by policy and funding and institutional priorities as much as by technological change."
"The limitations of AI tools — such as missing nuance, emotion and cultural context — reinforce the continued importance of genuine language learning for authentic communication."
In brief
- AI’s instant translation is challenging the future of language degrees, forcing universities to redefine the value of linguistic study in an increasingly automated world.
- Beyond basic translation, AI often misses cultural nuance and accuracy, risking a more English-centric world where deep intercultural understanding and critical thinking are eroded.
- Universities must prioritise intercultural competence over mere translation, investing in language programmes that develop the nuanced human understanding necessary for diplomacy, business, and global citizenship.
Across universities and workplaces, language barriers are dissolving rapidly. Generative AI tools can now produce fluent multilingual text and near-instant translation, enabling communication across unfamiliar languages.
This shift is raising fresh questions about the future of modern language degrees: if AI can translate, do universities still need to teach languages?
But this is not just about AI. The pressures on language learning in universities existed long before it — though patterns vary across institutions and systems — and may not be driven by AI alone, say global experts.
A decline before disruption
Across many higher education systems, language enrolments have been under pressure for some time. As Jen William, Professor of German at Purdue University and Co-Chair of the Modern Language Association’s Task Force on World Languages and Generative AI, points out, this is often linked to structural factors within education systems.
In the US, declining enrolments are linked to cuts in school-level language provision, leaving students less prepared to pursue languages at university. “Students are generally entering university with less language experience and thus there are fewer who are able, or who feel they would be able, to study a language as their major field of study,” she says.
In this context, AI may be less a cause than a convenient justification. William suggests it can be used “shortsightedly as an excuse or rationale” for cutting language programmes as a budget-saving measure, particularly in systems increasingly prioritising STEM disciplines. She adds that such decisions risk coming “to the detriment of developing well-rounded students with broad cultural understanding”.
These trends are shaped by broader system dynamics. Joseph Lo Bianco, Professor Emeritus, Language and Literacy, University of Melbourne in Australia, points to funding and policy settings as key drivers, particularly where humanities are deprioritised in favour of STEM and business programmes.
Even so, the picture is not uniform. Lo Bianco notes that while traditional degree pathways may be under pressure, there is also resilience in high-demand languages and in combined or non-degree forms of study.
Crucially, he emphasises that the overall decline in enrolments predates AI by many years. “The crisis in funding and system design still looks more powerful than AI in explaining the overall trend,” he tells QS Insights.
This suggests AI is entering an already shifting landscape shaped by policy and funding and institutional priorities as much as by technological change.
These trends are particularly visible in some English-speaking settings. In the US, Modern Language Association data shows enrolments in languages other than English fell 16.6 percent between 2016 and 2021, and have dropped 29.3 percent since their 2009 peak. In the UK, single-subject language degrees have declined most sharply, while combined and joint degrees have been more stable, according to British Academy analysis.
Australia presents a more mixed picture. A 2025 mapping study found modern language teaching remains “relatively vigorous” overall, but provision is concentrated in a small group of languages, with evidence that some universities have dropped particular languages or no longer offer full degree pathways.
However, these trends are not uniform, and some institutions and language programmes continue to show resilience.
The AI paradox: enabling and eroding
For Philipp Koehn, Professor at the US’ Johns Hopkins University whose work currently focuses on machine translation, AI’s impact is inherently dual.
Translation tools are now powerful enough to facilitate communication across languages, but they also introduce what he describes as “friction” and remain “limited to a cooperative environment”.
“As an optimistic take on language learning, such tools will provide an entry to communication among people who do not speak the same language, and the friction provides enough of an incentive to learn foreign languages,” he says.
He adds that these tools can also support language learning, alongside other digital tools that have recently become more prominent, such as spaced repetition apps.
However, Koehn cautions that “the technology also allows an easy way out: if relying on it is good enough, the effort involved in learning foreign languages may be seen as a much greater challenge without as clear benefits as before.”
This tension lies at the heart of the debate: AI can both expand access to communication and reduce the incentive to develop deeper linguistic skills.
What AI cannot replace
Despite rapid advances, experts emphasise that translation is not equivalent to language mastery.
Ana Niño, Senior Lecturer in Spanish at the University of Manchester in the UK, argues that language learning extends far beyond translation, fostering intercultural competence, critical thinking and a deeper understanding of societies.
“While generative AI can now translate text instantly and produce fluent writing across languages, universities still need robust language degrees and accessible language-learning opportunities,” she says.
The limitations of AI tools — such as missing nuance, emotion and cultural context — reinforce the continued importance of genuine language learning for authentic communication.
William echoes this shift in purpose. Language degrees today are less about producing translators and more about developing communicative and intercultural competence. “Someone with a language degree is able to identify bias and correct cultural stereotypes that are often an unfortunate feature of text generated by Large Language Models,” she notes.
These are the types of communication skills and cultural knowledge needed for an informed global citizenry, and that employers increasingly seek alongside technical expertise.
The idea that AI could replace language learning assumes translation is its primary purpose. But this overlooks how languages function in practice.
“Life doesn’t come with subtitles,” William says, pointing to the limits of relying on technology in high-stakes situations such as business negotiations or diplomacy.
AI systems also perform unevenly across languages, with lower accuracy for less widely used or digitally represented languages. “Languages that are less commonly spoken or not used as much on the internet have a much lower accuracy rate when translated by generative AI,” she adds.
A more English-centric world?
The implications extend beyond classrooms. Lo Bianco warns that declining language study could contribute to a more English-centric global system.
This creates asymmetry: English speakers remain largely monolingual, while non-English speakers are compelled to become multilingual. The result is not equality, but imbalance—along with risks of insularity and cultural misunderstanding.
Greater reliance on technology may further distance people from direct engagement with other cultures. While not inherently negative, Lo Bianco suggests it could erode “nuance and subtlety in our understanding of human life, rights and diversity”.
The future of language learning is unlikely to be defined by decline, but by adaptation.
AI has changed the context, not the value of language learning, highlighting its core purpose: not just translation, but understanding.
The question is no longer whether languages matter, but whether universities will continue to invest in them.