A few weeks ago, Microsoft released an interesting paper titled “Working with AI: Measuring the Occupational Implications of Generative AI” ranking professions based on their alignment with the current capabilities of artificial intelligence. Of all the groups, translators and interpreters were the most aligned. The paper (or to be more precise the screenshot of one chart, see picture) has circulated widely among translators and interpreters, and yet, most of the commentary around it reflects misunderstanding, denial, or outright dismissal.
Let me offer a few points that — while perhaps uncomfortable — deserve clear articulation.
- The paper is right in highlighting the strong alignment between AI and translation/interpreting.
Current AI systems are demonstrably capable of performing the core functions of translation and interpreting — even if imperfectly (humans perform them imperfectly too, but with a different degree of imperfection). The fact that these models can execute key linguistic tasks, sometimes in real time, at near-zero marginal cost, is not speculation but a measurable reality. The professions are among the top-ranked in the paper not because of any narrative bias, but because of a direct correlation between the tasks required and what AI can now do. - The structure of the profession makes it especially vulnerable
Translation and interpreting are highly specialized professions composed of narrowly defined tasks. A conference interpreter, for example, spends most of their time performing one activity: converting spoken content from one language to another in real time. The same applies to translators working with written texts. Unlike professions such as law or architecture, which involve a broad repertoire of extremely diverse and context-dependent activities, translation and interpreting are characterized by a near one-to-one mapping between occupation and task (the fact that those main activities can be further segmented in subtasks is insignificant here).
As the paper explains: “This score captures if there is non-trivial AI usage that successfully completes activities corresponding to significant portions of an occupation’s tasks.” This high correspondence between task and occupation is a vulnerability, and not just a theoretical one. It might explain why we’re already seeing a sharp decline in workload, fees, and general income stability for professional translators — and why, I am afraid, interpreters are likely next in line. - High AI alignment does not automatically mean immediate obsolescence
Yes, there’s a difference between AI alignment and becoming obsolete. Just because AI can perform a task doesn’t mean the associated profession vanishes overnight. It only says, as introduced above, that the vulnerability is high. When AI enables end users to perform those tasks themselves — as it increasingly does — not only does more of that task get done, but a smaller proportion is left for professionals. The impact is economic first, existential second. The emancipation of the user often translates into the marginalization of the expert. However, there are reasons outside the technology itself that makes obsolescence not an immediate consequence. Think of regulations, special use cases, quality expectations, trust, or human preference, to name just a few. - “Not-Us Syndrome” is alive and well.
As Richard Susskind noted in 2025, many professionals suffer from what he called the “Not-Us Syndrome” — the belief that automation can and will affect every profession except their own. Among the reactions to the Microsoft paper, this mindset is palpable. A significant share of the responses downplays the data, misinterprets the concept of task alignment, or dismisses the paper outright as irrelevant to their experience. - A mounting challenge for academia.
It is becoming increasingly difficult for universities to promote the professions of translators and interpreters, hence their curricula, with credibility, given the media coverage and empirical data like the Microsoft study or the Elis Survey. Many prospective students are already turning away, sensing a disconnect between institutional promises and market reality. Sure, they are changing their focus (some already cancelled the words translation and interpreting from their names), but remaining relevant will be a difficult challenge.
We must be serious when discussing the impact of AI on professions. For example this means developing the critical ability to engage with scientific studies for what they actually say — not accepting or rejecting them based on whether they support a comforting narrative. I’ve seen this firsthand: the same analysis can make you the hero of a group when it reflects their expectations, or the outcast when it doesn’t. But the responsibility of anyone discussing professional futures — whether in academia, industry, or public debate — is to go beyond slogans, personal interests, or fear of discomfort. The Microsoft paper has raised an important signal. The real question is: who is willing to look at it without turning away?
I worry that no one will step up — simply because no stakeholder group has anything to gain from doing so.
More than afraid, I am curious about the implications of machine T and I in society. Language professionals today could benefit a lot from data and trend analysis.
Since language is constantly evolving and we’re still within a small window before a drastic change in society, we should focus more on providing better insights. That way, translation and interpreting will remain relevant where human intervention is still needed.
Another possibility: learning rare language combinations!
Learning rare language combinations is indeed a viable (and to be honest even enriching) approach!
This is a stark and ominous reality for many to confront. But they won’t have to confront it alone.
I largely agree with your insightful analysis. However, I’d like to add one point: universities should continue to encourage the study of additional languages—ideally more than one. First, learning another language opens doors to new cultures, perspectives, and intellectual horizons. Second, it significantly enhances one’s employability across a wide range of industries, regardless of the profession. While training to become a translator may not be the most compelling reason for studying another language, there are many other valuable motivations worth emphasizing.
There is no doubt in my opinion that language learning is an will continue to be key, both a t personal and professional level. People want to communicate directly not through agents (human or ai interpreters) whenever this is possible. And language learning is the key for this.