Transformation of Approaches to Teaching Programming and Software Development under Influence of Generative AI

Abstract

This paper considers key transformational processes occurring in programming and software development training under the influence of generative artificial intelligence. Based on a series of interviews with junior students, a behavioral pattern leading to degrading skills of algorithmic thinking and problem decomposition is identified. A quantitative analysis of undergraduate theses from the Institute of Information Technology and Intelligent Systems (ITIS) at Kazan Federal University (KFU) over the past three years demonstrated a significant increase in AI usage in project work and the independent development of skills for utilizing LLMs in professional practice. The drawbacks of the uncontrolled use of generative AI in the educational process and the need to develop sustainable competence in using these tools in professional software development activities are shown. A three-tiered solution is proposed: improving assessment through analysis of the process and style of coding; regulating the use of LLM in project activities and final theses with an emphasis on assessing applicability of AI; and integrating AI tools into teacher work to update content and provide feedback. The proposed approaches have both conceptual and practiceoriented nature and aim to transform the process of training software developers in the context of the widespread adoption of AI tools in IT-industry.