1 Introduction

Discussed the application and impact of artificial intelligence technology, especially speech recognition technology, in the language service industry, particularly in the field of interpretation. Emphasis was placed on the potential of speech recognition technology in interpreting practice, including real-time transcription, memory assistance, and its popularization and application in the field of interpreting. At the same time, the reliability and technological maturity of speech recognition assisted manual interpretation were pointed out. The article also provides an overview of empirical research on speech recognition assisted interpretation from 2013 to 2023, both domestically and internationally, with the aim of summarizing research results and trends, and proposing suggestions for future development directions, providing reference for domestic interpretation technology research.

The Application of Speech Recognition Technology in Interpretation

Speech recognition technology (ASR/S2T) converts speech signals into text or instructions, which is the foundation of intelligent speech interaction. ChatGPT, a chatbot, uses speech recognition to achieve oral interaction and improve information retrieval efficiency. In the field of interpreting, the Verbmobile project has introduced speech recognition technology for the first time, and AIIC has listed it as a key research area. Speech recognition technology has become a typical application of the new generation of computer-assisted interpreting tools to assist in terminology recognition and digital interpretation. Technological progress will further improve the quality of interpretation, reduce cognitive load, and optimize the interpretation process.

Empirical Study on Speech Recognition Assisted Interpretation

Interpretation is a complex language processing task that faces challenges such as time pressure, cognitive load, bilingual activation, and multimodal input. The development of speech recognition technology has provided assistance for interpreting, and related research involves interpreting difficulties, quality, modes, and teaching. Gile's hypothesis of "problem triggers" provides a theoretical framework for the study of interpreting difficulties, including issues related to the speaker, source language, interpreter, and technology. Empirical research mainly focuses on vocabulary levels such as numbers and terminology, as well as issues related to spokespersons such as non-standard accents. Speech recognition technology can effectively reduce the cognitive pressure of interpreters and improve the quality and efficiency of interpretation. In terms of the accuracy of digital interpretation, speech recognition tools can significantly improve the accuracy of digital interpretation. In terms of automated terminology retrieval, speech recognition tools can optimize the way terminology is extracted during the interpretation process. In terms of non-standard accent recognition, speech recognition systems can help translators overcome accent barriers. In terms of interpreting quality assessment, there is no unified understanding of the impact of speech recognition technology on interpreting quality, but it has to some extent improved interpreting quality. In terms of the new hybrid interpreting work mode, speech recognition technology has expanded the application scope of interpreting technology and provided interpreters with more choices. In terms of interpreting teaching, speech recognition technology can enhance students' confidence in interpreting training and improve reflection efficiency, but its promotion and application in interpreting teaching still need to be strengthened.

Critical analysis of empirical research

This chapter delves into the key issues and limitations in empirical research on interpreting based on speech recognition technology, and proposes corresponding solutions. Firstly, the importance of the reliability and validity of measurement tools was emphasized, and it was suggested to increase the objectivity of effectiveness evaluation by strengthening theoretical construction, utilizing existing or self built scales, and integrating manual and machine scoring methods. Secondly, the issue of representativeness in sample selection was pointed out, and it was suggested to select subjects with small individual differences or adopt appropriate data collection and statistical methods to eliminate errors caused by small samples. Next, the variable manipulation of experimental materials was discussed, emphasizing the importance of controlling variables such as interpretation directionality, difficulty, and presentation style. In addition, the trade-off issue of ecological validity was explored, and the necessity of seeking a balance between experimental control and authenticity was pointed out. Finally, the selection of interpreting assistance tools was analyzed, and it was suggested that researchers face the limitations of current speech recognition technology in order to obtain more authentic, objective, reliable, and valid research data.

5 Future Trends and Prospects

The empirical research on speech recognition assisted interpretation needs to improve its research level and depth. Future research directions include human-computer interaction cognitive load, multimodal interpretation cognitive processing, and application-oriented research. The study of cognitive load in human-computer interaction focuses on the performance and cognitive load level of interpreters in different interactive tasks, exploring the influence of interactive interface elements on interpreters. The study explores the impact of multimodal information input on cognitive processing in interpreting, using empirical methods such as questionnaire surveys, eye tracking, and brain imaging. Application oriented research focuses on the teaching and practical application of speech recognition technology in the field of interpreting, analyzing market demand and communication effectiveness, and paying attention to ethical norms, privacy protection, and the definition of responsible parties. Empirical research on interpreting needs to comply with the requirements of the information technology era, promote industry university research cooperation, and realize social application value.

6 Conclusion

Emphasis was placed on the application prospects and challenges of speech recognition technology in the field of interpretation. The empirical research progress in the past decade was systematically analyzed, and improvement suggestions were proposed. The article aims to provide a comprehensive research overview for the academic community, promoting theoretical innovation and change. Future research should focus on interdisciplinary collaboration and innovative methods, while ensuring that technological developments serve to improve the quality and efficiency of interpreting.