Summary

Students answered the question, “Even though you can study a course on your own using generative AI, what makes Sungkyun Tutoring (study groups) better?” by explaining the advantages of learning with people.

· Communication and interaction through conversation help students check their understanding and expand their thinking.

· Learning contextual experiences and know-how from tutors and fellow tutees helps students set a clear direction for their learning.

· Explanations tailored to the characteristics of learners and learning materials are more intuitive and effective than those provided by generative AI.

-------------------------------------------------

   Generative AI is becoming an essential learning tool for today’s university students. From a student’s perspective, generative AI can quickly analyze complex information and provide well-organized answers. As a result, students may feel that when they do not understand something, they can solve the problem immediately with AI, without asking others or seeking advice.

   Sungkyun Tutoring, operated by the Center for Teaching and Learning Innovation at Sungkyunkwan University, is the university’s longest-running learning support program and has been continuously offered since the 2010 academic year. In Sungkyun Tutoring, tutors who have earned a grade of A or higher in a specific course study together with tutees who want academic support. They meet for more than 10 sessions and at least 15 hours over one semester. At its peak, the program had around 800 participants. However, since the rise of generative AI, participation has declined significantly, and in the second semester of the 2025 academic year, the number of participants is expected to be about 150. Even so, why do students still choose to join Sungkyun Tutoring study groups?

   In this Teaching Tip, we aim to explore learning experiences that cannot be replaced by AI by focusing on students who continue to participate in Sungkyun Tutoring, even in a time when AI provides so much information. In particular, we analyze the written responses of 64 students who participated as tutees to examine the appeal of “human-centered study groups” as recognized by university students in the AI era. Among the respondents, first-year students made up the largest group at about 50%. Many responses came from students in science and engineering fields, such as natural sciences, electrical and electronic engineering, and other engineering majors.

                                                                                                                                  Image source: ChatGPT (DALL·E), date created: December 12, 2025.

1. Communication and Interaction Based on Conversation

   When 64 tutees who participated in Sungkyun Tutoring were asked, “Even though you can study a course on your own using generative AI, what makes Sungkyun Tutoring better?”, the most common response (28.1%) was that it allows communication and interaction based on conversation. Students believed that even with the development of generative AI, real learning happens through sharing opinions and discussing ideas with others. Tools such as ChatGPT or Gemini can be useful, but there are still many parts that people can understand more clearly and intuitively through direct conversation. Students also felt that by sharing points of confusion with others, they could learn more than when studying alone.

   In particular, students felt that AI, which mainly communicates through text, has clear limitations. One student said, “Just as watching online lectures or reading books is different from attending an in-person class, learning through generative AI has limits because it is less effective in delivering understanding.” In other words, unlike text-only explanations, in-person settings allow students to listen, write, and communicate at the same time, while also understanding tone and nuance, which leads to better learning outcomes. Many students still believed that there is a large difference between AI and real people, and that face-to-face interaction makes asking questions easier and understanding clearer.

2. Sharing Contextual Learning Experiences

   The second most common response was “sharing contextual learning experiences” (23.4%). Students said that learning from tutors who had actually taken the course helped them understand which parts they should focus on. This was seen as a key advantage of Sungkyun Tutoring, because students study with senior students who have already completed the course. Since learning goals and evaluation methods can differ by instructor, information based on real class experience was especially helpful in understanding the course content. Students pointed out that this was a major strength that clearly distinguishes tutoring from learning with generative AI. Generative AI often explains all parts with similar weight and tends to imply that everything is equally important. In contrast, studying with a tutor helps students understand which topics and chapters are relatively more important.

   Beyond course content, some students noted that tutors connected what they learned in class to real industry situations and gave advice based on their overall university experience. As a result, tutoring was not just about delivering knowledge, but about experience-based learning that supported real student growth. Because tutors and tutees are also seniors and juniors, tutors shared their know-how with younger students, helped build personal networks, and talked about real concerns and experiences. Through this process, tutees felt they gained significant support.

3. Human-Centered, Personalized Support

   The third most common response from Sungkyun Tutoring tutees was “human-centered, personalized support” (17.2%). When studying with generative AI, students need to write very specific questions and clearly explain what they do not understand, which often takes a lot of time. In addition, when students are not even sure what they do not understand, it can be difficult to ask questions at all. Because generative AI mainly responds to the questions it is given, good answers depend on asking good questions.

In contrast, in tutoring sessions, tutors can explain content while considering the tutee’s level of understanding. When a tutee does not understand something, tutors can pause and provide more detailed explanations through ongoing questions and answers. Since tutors have already completed the course, they can easily understand which parts tutees are likely to find difficult. This allows tutors to adjust their explanations by balancing difficult and easier parts of the content. Through this process, tutees receive personalized coaching and can quickly improve areas where they are weak.

   Students also said that generative AI has difficulty understanding the context of questions and often takes more time to provide detailed explanations. In particular, AI explanations of visual materials such as graphs or circuit diagrams are hard to understand. There are also limits to using AI to analyze handwritten notes or problem-solving processes.

   Even now, when generative AI has become a common learning tool, students who participated in Sungkyun Tutoring continued to talk about learning experiences with people that cannot be replaced by AI. The key values identified in this analysis—active interaction through conversation, sharing contextual learning experiences, and human-centered personalized support—are all forms of learning that AI finds difficult to provide.

   Although the number of students participating in Sungkyun Tutoring has decreased, many students still feel that they learn more deeply by explaining ideas out loud, asking questions, and checking their understanding with others, rather than quickly finding answers alone through AI. Advice and learning guidance from senior students who have already experienced the course also provide a type of support that is qualitatively different from the balanced but uniform explanations given by AI. In addition, tutors can read a tutee’s current level of understanding, adjust explanations immediately, and provide customized feedback by identifying and addressing weak areas together.

   These results suggest what should become more important in teaching and learning in the AI era. If AI takes on the role of providing information, instructors, tutors, and peer learners should strengthen their roles in guiding learning and expanding students’ thinking. In the end, as AI continues to advance, the core value of learning becomes even clearer through interaction with other people. We hope this Teaching Tip helps instructors recognize the usefulness of AI while also better understanding the value of study groups, where students learn through cooperation, empathy, and shared understanding.

Sang-eun Lee, Min-young Ku, Ye-jin Kim (2025), Why Do Students Still Join Study Groups in the Age of AI? (CTL Teaching Tips #68). Seoul: Sungkyunkwan University, Center for Teaching and Learning Innovation.