EVALUATING USER TRUST AND PSYCHOLOGICAL IMPACT OF AUTOMATED SPEECH RECOGNITION (ASR) FEEDBACK ON NON-NATIVE ENGLISH LEARNERS IN PAKISTAN

Abstract

This research examines the link between speech technologies and human psychology by investigating the relationship between user trust and the psychological effects of feedback from an Automated Speech Recognition (ASR) tool for non-native English-speaking students in Pakistan. Using a quantitative correlational approach, the experiment involved a sample size of $N = 40$ college students enrolled in English linguistics at the undergraduate level, who were engaged with an accessible ASR tool based on the Whisper system from OpenAI. Psychological variation was determined through modified versions of the TAM model and FLCAS anxiety scale. Statistical analysis conducted in SPSS showed a significant positive relationship between the reliability of the ASR system and user trust ($r = .68$, $p < .01$). At the same time, frequent mistakes made by the ASR system caused speaking anxiety among low-level language learners.