CHATGPT IN FOREIGN LANGUAGE TEACHING AND ASSESSMENT: EXPLORING EFL INSTRUCTORS’ EXPERIENCE

The paper presents the results of the first stage of the research into the specifics of ChatGPT application in assessment development for English as a Foreign Language (EFL) courses. The article aims to investigate the perceptions of English language teachers regarding the utilization of ChatGPT in their teaching and assessment practices obtained through a survey in order to establish the degree of their familiarity with the possibilities of ChatGPT for teaching a foreign language and the mechanics of AI-generated test tasks design. The participants are mostly experienced English language instructors, affiliated with higher education institutions in Ukraine, the EU and the USA. Findings reveal varying levels of respondents’ confidence in ChatGPT's functionalities. A considerable proportion of educators appeared to be either unfamiliar with ChatGPT or possess limited knowledge of its capabilities. Hesitancy in implementing ChatGPT in teaching practice primarily stems from a lack of guidance as to its effective utilization closely followed by academic integrity concerns. However, educators who have employed ChatGPT recognize its potential as an assistant in lesson planning, content creation, task design, and assessment, especially for vocabulary, grammar, and writing. Despite acknowledging the tool's benefits, respondents express reservations about ChatGPT-generated tasks’ accuracy for assessment purposes. Challenges include the need for meticulous proofreading, issues with biased information, and insufficient task complexity. Nonetheless, overall satisfaction levels among educators implementing ChatGPT range from moderate to high, substantiating its value in foreign language teaching. Thus the survey stage of the research revealed the growing interest in ChatGPT’s potential for foreign language teaching and assessment among Ukrainian and international EFL and ESP instructors and underscored the need for guidelines and recommendations to create effective ChatGPT-generated test tasks. Future research will develop criteria for evaluating ChatGPT-generated tasks, provide prompt design guidelines, and formulate ethical policies for AI in assessment, addressing educators' concerns and