Do AI-generated field-specific vocabulary quizzes provide consistent results? Measuring learners’ performance and their perceptions

Artificial Intelligence (AI) is increasingly becoming an incredibly beneficial instrument for language education. There is a variety of AI-powered educational instruments that are embedded in miscellaneous platforms whether they are educational or non-educational. These AI-powered platforms are said to have the ability to improve English language learners’ communication skills through the provision of interactive and personalized learning experiences. Moreover, learning English for Specific Purposes (ESP) vocabulary is sometimes problematic for many learners and the search for new effective learning strategies is not an option anymore. In addition, there is a need to create and generate reliable, authentic, specific and valid ESP tests. Thus, the principal aim of the study is to examine the reliability of AI generated field-specific text-based activities and the attitude of cardiac care nurses towards these AI-generated activities. In this endeavour, four AI-composed quizzes were disseminated to cardiac care nurses (n=70) in various departments (e.g., Intensive Care Unit, Cardiac Care Unit, Adult Ward, Operating Room) at a cardiac care Centre and a questionnaire (based on a five-point Likert scale) was administered following the application of the quizzes to identify the attitude of the participants towards these quizzes. To confirm the results, interviews were also conducted with a subset of the questionnaire participants (n=8). The results revealed that the AI-composed vocabulary quizzes pleaded consistent results in terms of learners’ level of proficiency. Additionally, the nurses’ attitude was positive towards using these quizzes as an assessment tool along with raising their interest and motivation to learn more nursing vocabulary. Furthermore, a model for improving ESP vocabulary through field-specific, contextualized, authentic and level-suitable texts and activities has been constructed. The model is based on the findings of the questionnaire. Received on: 01 December 2024 Accepted on: 03 January 2025 Published on: 5 February 2025