ON THE QUALITY OF AI-GENERATED MULTIPLE-CHOICE QUESTIONS (ON THE MATERIAL OF TEACHING FOREIGN LANGUAGES)

Abstract

One of the current trends in modern methods of foreign language teaching is the neural-network-technologies-based development of linguodidactic materials. In this regard, the issue of how suitable such materials are for the use in teaching process acquires particular relevance. The article examines the quality of multiple-choice questions generated by artificial intelligence automatically at the request of the faculty member on the basis of a profession-related text. The main research method is a comparative analysis of the quality of question sets obtained from six digital platforms as a result of their pilot testing. Quality assessment is based on determining whether the generated tasks meet the requirements for their composition, content and language. The author identifies the key shortcomings of the generated questions. Taking them into account, a checklist designed to monitor and prepare the questions received for use in the teaching process is worked out.