Revealing Automatic Speech Recognition Technology in Simultaneous Interpreting: A Systematic Review

OpenAlex Turkish

This systematic review investigates the effectiveness of Automatic Speech Recognition (ASR) technology in simultaneous interpreting (SI), focusing specifically on its effects on interpreters’ performance and cognitive load. It highlights how ASR systems are employed in the field of interpreting and their potential influences on interpreters’ performance and cognitive load. This review adopted the PRISMA 2020 guidelines. Four databases (EBSCOhost, Web of Science, Scopus and Google Scholar) were searched, combining technology-oriented terms with simultaneous interpreting, performance and cognitive load. Inclusion criteria consisted of peer-reviewed articles in English which exclusively focused on the use of ASR technology both integrated into CAI tools such as InterpretBank and other tools such as Zoom, and the impact of ASR-supported SI on interpreters’ performance (e.g., accuracy) and cognitive load. Studies that did not meet the criteria were excluded. After eliminating duplicates and screening, fifteen studies met the eligibility criteria. The studies varied widely in terms of the experimental designs and methodologies, including error analysis (e.g. omission, additions etc.), eye-tracking, descriptive EVS analysis, interviews and subjective measures (e.g. the NASA-TLX and the Paas scale), and neurophysiological techniques (e.g. EEG and fNIRS) to investigate the relation between ASR and overall performance and cognitive load. Overall, the results suggest that ASR tends to enhance interpreting performance but the assumption is that it may also impose an additional cognitive load on interpreters. Nevertheless, according to neurophysiological data, such as EEG and fNIRS, ASR may alleviate cognitive load.