The purpose of the study is to examine the usefulness of mobile-assisted ASR dictation systems (Gboard, Siri or voice dictation on smartphones) for vowel pronunciation practice by looking at three aspects of its usefulness: pronunciation improvement by using ASR, accuracy of recognition, and the learners’ attitudes towards using this system. A list of 30 words containing minimal pairs of the contrasts /i/, /ɪ/; /æ/, /ɛ/; /u/, /ʊ/; /ɑ/, /ʌ/ and some distractors was given to 21 Macedonian EFL learners, divided into two groups, an experimental (n=11) and a control group (n=10). A mixed methods approach was used in this study. The quantitative part of the study included pre-test and post-test recordings which were transcribed by 10 native listeners to measure their accuracy gains, as well as a comparison between ASR written output of native speakers and that of non-native speakers, and another comparison between ASR written output of non-native speakers and human judgments. The qualitative analysis explored learners’ attitudes towards ASR by analyzing students’ Facebook posts throughout the practice period.

Findings showed that the experimental group improved their accuracy while the control group did not show any improvements. Next, the findings demonstrated a close relationship between ASR written output and human judgment with an acceptable agreement for most of the vowels. Nonetheless, ASR did not show high recognition of native speech, especially for the vowels /ʊ/ and /ʌ/. Qualitatively, the learners’ Facebook posts showed positive attitudes towards ASR. An occasional frustration with inaccurate feedback was also reported but learners generally enjoyed the training and found ASR to be practical and a safe environment for practice. This study recommends inclusion of mobile-assisted ASR in the EFL classrooms for raising students’ awareness of the vowel sounds in the English language with careful guidance from the teacher as well as focused and structured practice using individual words.