Abstract
The speech recognition technologies based on artificial intelligence (AI) are revolutionizing the way the second language is learned, especially in the process of enhancing the pronunciation of the English language among ESL students. This paper examines how AI-based speech recognition technologies can improve pronunciation accuracy, fluency, and confidence in Pakistani university students. The research design used was a mixed-method research design in which pre-test and post-test assessment as well as questionnaires and interviews were used. Participants used AI applications like ELSA Speak and Google Speech-to-Text during a four- to six-week time frame. Quantitative outcomes showed the improvement of the pronunciation accuracy significantly, and qualitative outcomes were expressed in the form of the increased learner motivation, autonomy, and engagement. Nonetheless, it was found that there were challenges like internet addiction, the inability to recognize accents, and access to technology. The paper finds that AI-based pronunciation devices are scalable, customized, and effective learning tools, and are therefore quite applicable in Pakistani ESL settings. When incorporated into language programs, these types of technologies can help to considerably increase training in pronunciation and speaking skills in general.