Bar-Ilan Closes Its Legendary Translation Department as AI Threatens More Professions

Bar-Ilan University announced last week that it will close its Department of Translation and Translation Studies ahead of the next academic year, and its researchers will be dispersed among other departments. The department was the only one in Israel that offered a comprehensive track including certification in translation, interpreter training and academic studies in translation research up to the doctoral level. Dr. Michal Shuster, a translator, language accessibility consultant and director of the language and cultural accessibility program at the Health Ministry, who was formerly a student and faculty member in the department, said, "It is a very highly regarded department in the field. The articles that came out of it and the doctors who graduated from it were recognized worldwide."

The closing of the translation program is seen as an early sign of the wave of so-called "dead professions" that artificial intelligence is expected to bring. Already today, AI can perform translation at a level that allows fairly good understanding of text in another language, and the expectation is that over time it will improve enough to translate professional documents, prose and even provide simultaneous interpreting at a satisfactory level for most clients. Uncertainty about the field's future is deterring students from choosing it as a career. "In recent years, only about three students a year, plus or minus, have enrolled in the department," said Prof. Amnon Albeck, the university's rector. Is translation really a dead profession? Are there certain types of translation, or certain components of the translation task, in which humans will always have an advantage? And which other professions are at risk?

At Bar-Ilan, they stress that despite the closure of the translation department, there is no intention to eliminate the jobs of translation researchers. "Demand for teaching translation has changed because the world has changed," Albeck said. "Demand for translators has dropped dramatically. Machine translation is not yet perfect, but it is approaching being good enough. Will that happen in two months? In two years? In any case, we understand the direction." He said there may still be a need to check AI translations. "But who needs to check them? Should it be a translator? Or rather a professional in the field?" Albeck emphasized that while the profession itself is changing, translation studies remain a significant and living field. Therefore, the department's researchers will not be replaced, but will be absorbed into other humanities departments. In his words, "It will always be interesting to study the connection between translation and the society in which we live. One can study Bialik's translations, and also those produced and to be produced by the machine."

As an example of change in academia, he pointed to the library science department that was closed in the past. "They taught how to catalog, where to put a book on the shelf. Those are skills that no longer need to be taught at the university." And still, there are many librarians. "True, and the information profession did not disappear, it became computerized information science." Are there other departments at risk of closure? "Not in the near future, but we are an adaptive university that must deal with reality," Albeck replied. "For example, next year we will open the program in 'Computational Artificial Intelligence,' a track born against the backdrop of declining enrollment in programming departments, given the assessment that AI will be able to program at a human level. That does not mean that computer science is irrelevant."

From translating words to translating messages

So will the translation profession survive, even in changed form? Shuster believes so: "Translation is not a technical action, but a cognitive one. Given the place that the interface between languages occupies in our world, there is a great deal to learn and research about translation, and also a future professional path for translators themselves." At the most recent meeting of the Translators Association, it emerged that the added value of the future translator will lie less in language knowledge and more in culture. "The translator becomes a localization expert," Shuster explained. "He or she will know how to formulate the message in a way that the target audience connects with most, and make sure the wording is in the right register."

According to her, the translator will know how to choose words that bring closer and fit the linguistic style and accepted manners of the target audience. "The understanding is that translation is the transfer of messages from language to language, not of words."

Will the machine not learn to do that too? "Cultural adaptation has a patterned, fixed part that the machine can perform," Shuster said, "but part of it depends very much on the situation, on a specific platform, and requires very human sensitivity."

Shuster studies the connection between translation and the realization of rights, and applies it, among other things, in the health system. According to her, the link between language and rights is also essential in legal systems, bureaucracy and workplaces: "Here, the interpreter has a unique role because he enters an arena with built-in inequality." She explains that this is usually a situation in which the service recipient does not speak the majority language and lacks knowledge, while the service provider holds the power to grant or deny what is wanted. The Health Ministry, for example, operates translation services as well as cultural mediators, combining translation with mediation. "The machine's ability to do this is limited," Shuster explained. "Unsupervised machine translation may widen social gaps. A human interpreter from the culture of the minority-language speaker inspires trust, ensures understanding and raises a flag when needed, things a machine is not yet capable of."

Shuster added that even in technical fields today, machines are prone to critical errors: "The meaning of a mistake can be significant, so we still do not see humans removed from the process in places where the cost of an error matters. In addition, humans are needed to train the machines for different roles."

On the department's closure, Shuster added: "My role as director of language and cultural accessibility in the health system could not have existed without the research background I received at Bar-Ilan, and you can see the difference between systems that do not rely on professionals."

Even studying glazing will not help you

Dr. Roy Tszena, a futures researcher at the Blavatnik Interdisciplinary Cyber Research Center at Tel Aviv University, believes that in the future the machine will be able to imitate the human brain in all its aspects, including cultural adaptation. "The human brain is a machine," he explains. "And if it is a machine, then at some point a machine will come along that can imitate it. The question is when."

According to him, despite machines dominating the field, human translators will continue to be needed because of the personal touch. "People value what is scarce," he says. "When machine translation becomes cheap and accessible, I believe that for special events they will call in a human interpreter, who will provide prestige value."

He compares this to a virtuoso musician in an orchestra, emphasizing the fact that he himself, as a human being, plays from his physical body and emotional world, similar to a translator who will use gestures or jokes to highlight the human dimension. "Will his translation be better? Not necessarily," Tszena says. "Maybe it will be a performance, but one that sells. Even in prose, while a machine will provide an accurate and enjoyable translation, the human translator will be able to give it a personal nuance."

Tszena emphasizes that this future scenario, machine translation with human oversight, is still better than the machine alone. "The same is true in programming. At Anthropic, which develops the Claude language model, 90 percent of the code is written by code itself, but humans direct and supervise it. Output increases eightfold that way. So I do not agree that programming is a dead profession; it makes sense to study computer science in order to work alongside artificial intelligence at the highest level."

However, Tszena warns that manual trades are not immune either: "We are heading toward a physical robot revolution, which will be able to work in glazing and plumbing and even take the dog for a walk. This will not happen in the coming year, but it could happen within our lifetime."