A sworn translator is not just someone who converts the wording from one language to another. They are the guarantor of legal truth. Every certification of a translation is a public document. Any error can give rise to a criminal offence, civil liability or result in international litigation. There are no shortcuts: words carry weight, and human judgement carries more weight than any algorithm can.

Imagine the scene in the courtroom. The judge looks up, but instead of the sworn translator, is confronted by an instrument of artificial intelligence. There is a cold light and a tense silence. The software, in the flat voice of a virtual assistant, solemnly declares: “This translation is faithful to the original”. Everyone holds their breath. Who is it that is in a position to say whether the assertion is false? No one. No one that is but the human translator who pressed ‘send’ and swore to its accuracy, fully aware of the weight of their words.

And there’s the rub.  Would we ever entrust artificial intelligence with the task of translating and certifying a billion-dollar contract? Would we have it interpret a Will written in Sanskrit for an international inheritance? Worse still, would we trust its translation of a diplomatic accord between two countries? AI may now be endowed with remarkable linguistic precision, but it lacks a true sense of context, an understanding of the legal implications, or any cultural awareness. Translating from English is complicated, while translating for example from Sinhalese or Bulgarian, with a plethora of technical terms and specific legal provisions, requires expertise, intuition and experience. Such skills cannot at this time be truly replicated by any algorithm.

To passively rely on software is to play dice with the law. Every error generated by the machine becomes the responsibility of the translator who certifies it. There is no plea available that ‘the algorithm got it wrong’. The oath remains wholly human, and the responsibility is human. AI can be an invisible trap, tempting those seeking efficiency at the expense of prudence.

Picture some other scenarios. AI certifies a contract for the sale of a nuclear power station, while the translator looks on in horror; it translates a merger between multinationals without picking up on a clause that nullifies billions of dollars’ worth of rights; it interprets a will in a way that misses a nuance, that turns the whole inheritance process into an international dispute. They are extreme cases, but they are not science fiction. Certainly, the technology exists, and yes, it is sophisticated, but should we really trust it?

AI can be an aid, a co-pilot, or a tool to speed up the work. It cannot however be a substitute for the critical judgement of a sworn translator, nor can it be held criminally or civilly liability. Any apparently minor terminological error can have substantial consequences, whether they take the form of litigation, the voiding of clauses, or of financial loss. The translator still plays an essential role in distinguishing between legal precision and a mere combination of words.

In the final analysis, the sworn translator plays the role of sole guarantor of the fidelity and accuracy of the translation. The algorithm is a tool, powerful but lifeless. Every professional should pause to ask themselves whether they really want to run the risk of entrusting legal certainty to a machine incapable of a moral conscience, prudence and responsibility?

This is no mere technical challenge, but rather an ethical, professional, or even philosophical one. The illusion of digital efficiency must not be allowed to cloud our sense of responsibility. While AI makes its virtual declarations, the human translator knows that, in the end, only the human translator themselves can swear, understand and answer for the consequences. In short, it is only a human translator that can stand before the court and say: ‘I stand by these words’.