This study explores the differences in the student-AI interaction (SAI) process among students with varying attitudes towards AI and drawing skills during a public advertisement drawing task. The research is conducted using think-aloud protocols from 20 Korean undergraduate students, who are categorized into four groups based on their attitudes towards AI (positive or negative) and drawing skills (high or low). The study employs lag sequential analysis to identify statistically significant linear patterns and chronologically aligns these patterns to understand the overall SAI process. The findings reveal distinct differences in the SAI processes among the groups, highlighting the importance of considering students' characteristics and AI design in educational settings. The study suggests that educational AI should be designed to enhance metacognition and emotional engagement, and that a diverse team of stakeholders should collaborate to develop AI systems. Implications for practice and policy are discussed, emphasizing the need for holistic AI integration, increased metacognitive activities, and the development of explainable AI to facilitate effective student-AI collaboration.This study explores the differences in the student-AI interaction (SAI) process among students with varying attitudes towards AI and drawing skills during a public advertisement drawing task. The research is conducted using think-aloud protocols from 20 Korean undergraduate students, who are categorized into four groups based on their attitudes towards AI (positive or negative) and drawing skills (high or low). The study employs lag sequential analysis to identify statistically significant linear patterns and chronologically aligns these patterns to understand the overall SAI process. The findings reveal distinct differences in the SAI processes among the groups, highlighting the importance of considering students' characteristics and AI design in educational settings. The study suggests that educational AI should be designed to enhance metacognition and emotional engagement, and that a diverse team of stakeholders should collaborate to develop AI systems. Implications for practice and policy are discussed, emphasizing the need for holistic AI integration, increased metacognitive activities, and the development of explainable AI to facilitate effective student-AI collaboration.