Classification of problems that have manifested themselves in the training of students after opening up access to artificial intelligence services

The relevance of the study arises from the rapid proliferation of artificial intelligence (AI) services in education, which creates new challenges for pedagogical practice. The use of AI by students leads to decreased motivation for independent learning, problems with academic integrity, deformation of critical thinking and the need to adapt teaching methods. The objective of this paper was to categorise the key problems arising from the use of AI in education and to propose solutions, including the author's pedagogical model for teaching programming. The research methods included analysis of scientific publications, experiment with control and experimental groups of students, quantitative and qualitative analysis of the results, questionnaire survey, content analysis of code and comparative analysis of traditional and AI-integrated approaches. As a result of the study, a classification of problems including decreased motivation, lack of AI skills, violations of academic integrity, deformation of critical thinking, the need to adapt teaching methods, changes in knowledge assessment and digital inequality was performed. A pedagogical model for teaching Java programming based on individualisation, project structure and AI integration was proposed. The model included setting up personal projects, step-by-step mastering of the material, code performance control and formative assessment. The experiment showed that the application of the model increased the amount of code executed by 216%, the average score by 18.3%, and the students' motivation by 36%. At the same time, the level of independence decreased slightly (–1.35%). It was found that the use of AI in teaching requires a revision of control methods, including oral defences, portfolios and cross-checking of work. The practical value of the paper lies in the development of an adaptive teaching methodology that minimises the negative impact of AI and promotes programming skills. The model can be scaled to other disciplines that require practical application of AI. The results of the study are useful for teachers, methodologists and developers of educational programmes