Subtitling
Article reserved for subscribers
Errors or eccentric misinterpretations in the subtitles of films and series increasingly betray the use of generative artificial intelligence. An alarming prospect for translators, who already work in degraded conditions.
It’s the beginning of an American film or series, two characters are having a dialogue. The subtitles suggest that they know each other since they speak informally. However, from the next scene, they adopt the informal address before, a little later, using familiarity again. The effect is disturbing, even horrifying, but unintentional since it is undoubtedly a translation carried out by a machine and not by a professional. With the expansion of applications fueled by generative artificial intelligence and the enthusiasm it unleashes in the tech world, this kind of inconsistency likely to ruin a scene, or even the entire film, risks multiplying.
The threat has even become clearer in numerical terms since the publication, on December 4, of a survey carried out by the company PMP Strategy on behalf of the International Confederation of Societies of Authors and Composers (Cisac), which brings together more than 220 authors’ societies in 116 countries. According to this study, the market for musical and audiovisual content generated by AI will experience, in the next five years, “exponential growth”, increasing to an estimated value of 64 billion euros instead of the current 3 billion euros. A revolution therefore, which would take place to the detriment of creators who would suffer
