Human-AI Interaction: Research and Innovation in Real-Time Speech Generation, Interpretation, and Translation
摘要
The field of Computer-Assisted Interpreting (CAI) has advanced with tools such as InterpretBank and InterpreterAssist, designed to optimize interpreters’ workflows through technologies like Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) synthesis. However, their adoption faces challenges due to the limited generalization of studies, the perception of increased cognitive load, and the lack of integration of functionalities into a single tool. A recent project, which is currently being carried out by the authors, proposes the development of a comprehensive and ergonomic CAI tool that integrates advanced technologies such as Artificial Intelligence (AI), ASR, TTS, Neural Machine Translation (NMT), and Natural Language Processing (NLP). This approach aims to overcome economic barriers by offering an accessible, user-centered product. Additionally, Learning Analytics and Eye Tracking will be employed to evaluate its impact, measure cognitive load, and iteratively improve its design. As part of this initiative, the INNOVATRAD project is proposed, running from the 1st of January 2025 to 31st of December 2027 and founded by the regional government of Madrid (Spain) through a competitive program. The project is divided into three phases: creating a demo version, stabilizing the tool, and launching an optimized version. Continuous feedback from interpreters will be prioritized to ensure the tool’s practical utility, fostering its adoption in educational and professional settings. This approach promises to revolutionize interpretation by integrating functionality, accessibility, and user-centered design to transform professional practice.