WILL AI REPLACE TRANSLATORS?
Claret Christensen traces the translation’s journey from an ancient bridge between cultures to an AI-assisted industry where experienced linguists face falling rates, insecure work and a troubling contradiction.
By Claret Soares Christensen.
Technologies have historically been agents of transformation in society. They enable the development of new models of work and life and sometimes promote the transformation or the accelerated elimination of certain professions. Recently, a Microsoft article listed some professions at risk of disappearing due to AI use. Translation is one of the professions that has been significantly affected by the use of Large Language Models (LLMs) in the field, both for technical and literary translation.
As an ancient profession used in various contexts to connect people with different languages and cultures, it has been discussed using different approaches throughout history. Translation was essential for the early dissemination of key cultural and religious texts and concepts.
Discussions about the role of translation have occupied intellectuals and philosophical spheres since earlier centuries, for instance Cícero and Horace in the 1st century and St. Jerome in the 4th century. Later in the XX century, and more precisely in its last decades, translation became a distinct field of study. The reflection on the processes, techniques, and resources shed light on the profession in the Academic field.
While many professionals worked their entire lives with basic dictionaries, grammar, and technical glossaries to solve their translation projects, technologies used in this field advanced very quickly.
Translation occupied an important role in the globalisation of the markets. It was necessary to create a bridge between countries and cultures, considering the values of the target receptor, as we could call it in communication. Translation Studies was introduced at universities with different approaches. Translators were assigned to different groups based on their preferences for technology use since their university years, while some literary translators were defenders of the traditional way of translation, working at a slow pace and without using technologies such as Computer Assisted Translation (CAT) tools.
The technical translators, on the other hand, could see the benefit of adopting CAT tools, as they could create their own memory banks and reuse them in their next projects. It was an accelerator for productivity in technical translation. CATs are the translator's helper. The tool that ensures consistency in terms and style, thereby contributing to the quality of the translation.
CAT tools are required by many companies when they contract a technical translation project. They choose their favourite programs, and some companies have their own CAT, fed with their specific or restricted terms, memories and glossaries.
In July 2019, a major Danish publishing company launched a project to translate Danish literature into 20 languages. There were books in different genres, for different readers, from children’s collections to adult publications. It was an important project that invested in audiobooks as the future of literature. At least 20 translators were hired under a contract with a good volume of texts and fair conditions. Then 2020 began, and the world was forced into lockdown because of the COVID-19 pandemic. The demand for medical content translation and localisation increased dramatically worldwide. The technical translators were busy, and their work was well valued. The Danish project was then readjusted to reduce the number of languages and focus on just 5. Yet, as a result of the global demand, the localisation and interpretation programs and apps emerged and were improved.
When the world reopened its borders, that same demand for translation services changed again. The big projects for Large Language Models became available, and linguists were called upon to evaluate the quality of their outputs.
In the beginning, there were many tasks, and the quality of the answers needed improvement. The process was very fast, and a few months later, the quality of their output improved, becoming acceptable or, in some cases, surprisingly good. Then the demand for these translation projects declined rapidly, though linguists still occasionally receive projects that were translated using AI for revision or validation. The translation rates were reduced to values that were sometimes unacceptable. Many professionals began to transition careers.
The LLMs that the linguists helped to improve now do their job. The most recent data about the language industry, the NIMDZI 2026 report, confirms our conclusion in this short article. Its State of Language Industry 2026 publication says:
“On the provider side, this technological shift has fundamentally altered company dynamics and human resource requirements. Traditional in-house linguistic and administrative project management roles are being downsized or repurposed as automation increases productivity threefold. In their place, there is a surging demand for computational linguists, AI business developers, and prompt engineers who can navigate both language and machines.”
The report also states that in the past 2 years, a number of linguists have transitioned in their careers:
“For the linguists themselves, this transition has been polarising; experienced specialists who adapt to auditing, cultural editing, and managing AI workflows are commanding higher premiums, while those relegated to basic, unrewarding post-editing face significant downward pricing pressure and job dissatisfaction. This dynamic has led to a noticeable talent exodus, raising widespread industry concerns about a looming shortage of highly qualified professionals capable of authenticating AI output.”
The industry wants highly qualified professionals but at the lowest rates and the longest payment terms. The sum doesn't add up. The shortage of well-qualified professionals is a consequence of that.
