Abstract:

The article is oriented to the use of artificial intelligence in teaching programming and solving programming tasks. Studying with the support of artificial intelligence tools in combination with other methods offers effective and efficient acquisition of algorithmization and programming. The method from the natural learning model to machine learning and from the use of machine learning principles to control the natural learning process, which leads to the acquisition of new knowledge, is generally applied. In our research, our aim is not to have AI solve our tasks and problems for us, but to help and advise us in solving problems in a certain area. It is about supporting and not suppressing your own ideas and creativity.

We tested several artificial intelligence tools for generating program code for solving selected tasks. We decided to evaluate the obtained solutions (resulting source code) with code generators and chatbots. For deeper analysis and comparison of performance results, we chose 3 chatbots: NLP Cloud Playground Chatbot, BardAl and ChatGPT and two code generators: CodePal, NLP Cloud Playground for code generation. When researching and evaluating the results, we focused on finding answers to the following questions. "Can AI code generation tools generate code of adequate quality?" and "Does the communication language affect generation speed?" We concluded that every single AI tool can generate adequate code, but it may not be the optimal. The language of communication itself can have some influence on the speed of obtaining a solution.

As for the practical use of AI in the teaching process – tools of AI should not be banned by directive, but both teachers and students should use them as a "consultant" and co-researcher with a large dose of critical thinking and rigorous evaluation of (jointly) obtained solutions.

Keywords:

AI, AI tools, Programming teaching and learning, chatbot, code generator.