Comparative Analysis of Pattern Tasks Experiences Between Fifth Grade Secondary School Students and ChatGPT Artificial Intelligence Tool

<p><span>The technology that rapidly changes and advances the world necessitates interaction in all areas of life. Considering the role of technology in education today, the depth of this interaction offers various education opportunities. One of these opportunities is the technological tools used in educational settings. Technological tools, which add different dimensions to education, are quite significant. In this context, the use of artificial intelligence tools in educational environments in recent years is noteworthy. AI chatbots are particularly prevalent in education. Students and teachers can interact with these chatbots in written or spoken form. ChatGPT, developed by OpenAI, is one of the most frequently used chatbots recently. It can generate texts similar to those constructed by humans, understand users' language, answer their questions, and perform certain tasks. In addition to being a large language model, ChatGPT is also used in various disciplines, including mathematics. Given the importance of patterns in algebra teaching, it is inevitable that children encounter pattern tasks at an early age. It is important to reveal the similarities and differences between the answers provided by students and those generated by the ChatGPT AI tool to pattern tasks. Accordingly, this study examines the differences between the answers given to pattern tasks by fifth-grade students and the ChatGPT AI tool. A qualitative approach using a case study design was employed in line with this aim. The study was limited to fifth-grade students and the ChatGPT AI tool, with their responses to pattern tasks examined in detail. Purposeful sampling was used to select eight fifth-grade students and the ChatGPT AI tool as the units of analysis. Criterion sampling, a type of purposeful sampling, was utilized, with voluntariness and being in the fifth grade considered as criteria for participant selection. The participants comprised 100 fifth-grade students studying at a middle school in Istanbul and the ChatGPT AI tool. Pattern worksheets and screenshots from the ChatGPT AI tool were used as data collection tools. The pattern worksheet consisted of four number pattern tasks with constant increasing and decreasing patterns and 16 sub-problems. The researcher developed the pattern tasks using achievements from the mathematics curriculum and literature to ensure content validity. The pattern worksheet was later administered to students participating in clinical interviews with additional questions for in-depth examination. Ethical committee approval and necessary permissions from the provincial and national education directorates were obtained before data collection. Clinical interviews were used as the data collection method. Initially, the pattern worksheets were administered to the study group, and students were provided with the necessary information about the study before completing the worksheets. Students were given 80 minutes to answer the pattern tasks. Based on their responses to the pattern worksheet, students were categorized. To obtain rich data for comparison with the ChatGPT AI tool, clinical interviews were conducted with students considered to represent each category best. Data analysis was performed using content analysis. According to the study findings, students most frequently used random arithmetic operations without continuing the pattern to reach the close terms of the pattern tasks. The ChatGPT AI tool, however, utilized operations without continuing the pattern or writing formulas to reach the close terms of the pattern tasks. For the middle terms of the pattern tasks, students again predominantly used random arithmetic operations without continuing the pattern, whereas the ChatGPT AI tool used operations without continuing the pattern, writing formulas, or using formulas. Students mostly relied on random arithmetic operations to reach the distant terms of the pattern tasks. At the same time, the ChatGPT AI tool used operations without continuing the pattern, writing formulas, or using formulas. In inverse problems of the pattern tasks, students primarily used random arithmetic operations, while the ChatGPT AI tool employed formulas without continuing the pattern. Students faced difficulties and left blank responses, particularly regarding the pattern task's distant terms and inverse problems. Additionally, the constantly decreasing pattern task was the most challenging and frequently left-blank pattern task for students. Based on the responses provided by the ChatGPT artificial intelligence tool, students can consult this chatbot with any problems they encounter regarding pattern tasks. Although this robot often arrives at the correct answer, it can occasionally provide incorrect responses. Therefore, students need to use the ChatGPT artificial intelligence tool under the supervision of a teacher during the instructional process. Additionally, ChatGPT can present errors related to computer programming languages and Turkish character usage in its solutions. These are some of ChatGPT's limitations. This study is limited to fifth-grade middle school students and the topic of patterns. In future studies, similar research can be repeated with different grade levels or subject areas to determine how ChatGPT can be utilized in students' instructional processes.</span></p>