My job as an editor and translator has involved helping authors refine manuscripts, suggesting ways to reduce repetition, rewriting awkward sentences or passages, and highlighting areas that could confuse the reader. These tasks have always required insight, care, and placing myself in the reader’s shoes.
Today, as I’m increasingly tasked with editing machine translations, my job is not just to interpret the author’s message but also to check how well the AI has interpreted it. Has a subtle point been misinterpreted? Has the AI flattened the author’s voice? Does the AI sound fluent and confident, yet is actually wrong? Because machines, like us, do get it wrong at times, and what is striking about machine translations is that while some parts are incredibly accurate, other parts completely miss the mark.
The role of editors and translators today has shifted to become more like quality gatekeepers and AI overseers. In fact, our role increasingly lies not in fixing sentences but in ensuring meaning survives the machine. More importantly, it involves safeguarding the subtlety and depth of the human voice because however good these AIs are, they are simply too perfect and too smooth. And human beings are anything but that.
The day will likely come when an AI will have the ability to write a paper and oversee its quality. It may even be capable of actually conducting research. By then, the role of editors will probably have seen another transition, shifting from supervising the AI to training it and guiding its evolution. Giving it something we may no longer have: our human voice.