Speakers
Description
ChatGPT has already revolutionized our ability to have conversations with a computer. At present this is generally limited to typing rather than speech itself. Utilization of speech synthesis and recognition (SR) technology for English speaking practices could present alternative methods to assist in the development of speaking and listening skills. However, there are still challenges when utilizing such technologies in the classroom. In human conversation we are acutely aware of visual and auditory clues that indicate when it is our turn to speak, and at present, this technology is not present within AI models, and thus presents a significant challenge to the learner utilizing SR technology. Moreover, native speakers can recognize when ChatGPT goes 'off script' while this can be challenging for L2 learners. This presentation will elucidate how we addressed these issues and how effective they were. First, we developed a system that integrates SR and ChatGPT allowing English learners more opportunities to practice speaking and listening skills in a customizable framework based on the course needs and student needs. We then recruited L1 and L2 participants from various countries who interacted with our ChatGPT-SR speaking program to determine the feasibility of using such technology in the classroom, what difficulties students encountered, and actionable techniques that instructors can apply in their own classroom.
| Keywords | ChatGPT, AI, speech recognition, prompts |
|---|
