The study of automatic interpretation based on artificial intelligence (AI), using the example of the Wordly platform

The article explores automatic interpretation based on artificial intelligence (AI) technologies, using the example of the Wordly platform. With the development of neural networks and AI technologies, automated interpretation systems are becoming more and more in demand, but their quality and effectiveness remain insufficiently studied. Due to the relative novelty of this technology, the scientific literature contains a limited number of fundamental studies devoted to assessing the quality of automatic interpretation, which determines the relevance of further analysis of such systems. As part of the research, an experiment was conducted aimed at analyzing the quality of translation provided by the Wordly platform. A fragment of an article from the newspaper Kommersant was used as the source material, which was voiced by the author and processed by an automatic interpretation system. The results were analyzed in detail in order to identify the features of the platform and evaluate the accuracy of the translation. The methodological basis of the research includes conducting a practical experiment, comparative analysis, as well as the use of observation and generalization methods. The scientific novelty of the research lies in the appeal to a little-studied area of translation theory and practice – automated interpretation systems. The results obtained can contribute to the further study and improvement of such technologies, as well as to the development of methods for their assessment and application in professional translation activities. An analysis of the results of the experiment conducted using the Wordly platform allows us to conclude that, despite the active development of artificial intelligence technologies and automatic interpretation, the quality of translation carried out by this system remains insufficiently high. Errors are observed at all stages of the multi-stage process, including speech recognition, audio signal conversion to text, and subsequent translation generation. In addition, a significant drawback was the high time delay between the perception of the original utterance and the reproduction of the translated text, which averaged 10 seconds. This indicator significantly exceeds the standard 2-3 seconds allowed in simultaneous interpretation, which reduces the effectiveness of using automatic interpretation in the professional field.