The Center for Teaching and Learning Innovation has reviewed the trends in students' use of generative AI through Teaching Tip No. 42 (How Are Our University's Students' Experiences with Generative AI Changing?). This review is based on three surveys conducted from June 2023 to June 2024. The results show that students are increasingly using generative AI actively, and their satisfaction and positive perception of it are also rising. So, how are students using generative AI in university, and how do they want to learn about it? In this Teaching Tip, we will explore the questions students have regarding generative AI and discuss how professors might consider utilizing generative AI in their classes.

1. How Students Mainly Use Generative AI for Learning

  As the use of generative AI continues to grow, it is important to understand the main purposes for which students utilize AI in their studies. By comparing responses from the second semester of last year (November 2023) and the first semester of this year (June 2024), we can see that the most common purposes for which students use generative AI are "general research" and "programming and coding". The percentage of students who reported using generative AI for "general research" was 71.1% (295 students) in the second semester of last year and 62.6% (265 students) in the first semester of this year, making it the highest proportion, but it showed a decrease of 8.5 percentage points.  Additionally, "programming and coding" was used by 52.5% (218 students) in the second semester of last year and 52.7% (218 students) in the first semester of this year, with more than half of the students responding in both semesters and showing little difference in response rates between the two semesters.

  On the other hand, the use of generative AI for "summarizing specific materials (such as papers, articles, books, web pages)" and "structuring outlines for submitted work (essays, reports, PPTs)" increased significantly in the first semester of this year compared to the previous year. For "summarizing specific materials (such as papers, articles, books, web pages)," 33% (137 students) responded in the second semester of last year, but this increased by 17.4 percentage points to 50.4% (213 students) in the first semester of this year, indicating that over half of the students now use AI for summarization. Also, the use of AI for "structuring outlines for submitted work" increased by 13 points to 37.1% (157 students) in the first semester of this year. Additionally, the percentage of students using AI for "writing submitted work (essays, reports, etc.)" increased by 4.9 points. These trends suggest that students are now using generative AI for more specific and diverse academic purposes. Initially, students primarily used AI for general queries, much like an internet search. Now, they are leveraging AI to understand summarized versions of study materials for efficient learning and to structure outlines for their assignments, indicating a shift towards more targeted and practical applications of generative AI in their studies.

2. What Students Expect the University to Provide Regarding the Use of Generative AI

  Students primarily expect guidance from the university on "how to effectively use generative AI for learning." This was responded by over 70% of respondents in both June 2024 (71.6%) and November 2023 (70.4%). The second most desired guidance is on the "practical use of generative AI in specific fields (such as learning, research, coding, and competitions)," with around 65% of students indicating this in their responses. This demonstrates that students want hands-on experience and examples of how to use AI in practical areas they can encounter in university. Notably, the interest in practical applications of AI increased by 3.7 points from last year's second semester to this year's first semester, indicating a growing expectation for AI to provide tangible benefits in their academic pursuits. On the other hand, interest in "ethics of using generative AI" decreased by 7.2 points compared to the previous semester. This suggests that, in addition to meeting the demands for "how to effectively use generative AI for learning" and "practical applications of generative AI in various fields (learning, research, coding, competitions, etc.)", it is important to uphold the "ethics of using generative AI," even though students' interest in this area is declining.

3. “I Want to Learn by Using It in Regular Classes”: Preferred Methods for Learning About Generative AI

  In the first semester of 2024, a survey included a question about students' preferred methods for learning about generative AI. The survey provided multiple-choice options such as learning "through regular courses", "from peers and friends", "through extracurricular programs organized by their affliated department colleage", "with department peers and seniors/juniors", "with club seniors and juniors", "through extracurricular programs organized by specialized centers," and "other," allowing respondents to select multiple answers.

The results showed that the most preferred method was "through regular courses" with 36.4% of students choosing this option. The next most popular choice was the informal learning method "from peers and friends," selected by 29.3% of the respondents. Meanwhile, opportunities to learn through "extracurricular programs organized by their affliated department colleage" which offer learning divided by academic fields, were chosen by 11.3% of students. Learning "with department peers and seniors/juniors" was selected by 9.5%, and "with club seniors and juniors," focusing on interest areas, was chosen by 7.1%. This indicates that only around 10% of students preferred learning within existing communities.

  Some students provided specific answers under the "other" category. They mentioned preferring to learn about generative AI independently through YouTube or video content, or through AI study groups. Additionally, some students responded that they "prefer to use generative AI as a tool in existing major courses rather than creating new courses to learn about generative AI," and for they want the university to support the use of generative AI for it.

4. "How Do Professors Use ChatGPT?": Questions Students Want to Ask Professors About Using Generative AI

  We also explored what students are curious about regarding the application of generative AI in regular courses. When asked what they would like to ask their professors about the use of generative AI, 350 out of 423 respondents (83%) provided responses.

  Using ChatGPT, we analyzed the subjective responses and categorized them into eight types, as shown in the graph below. The largest category was "How to use generative AI (35.5%)“, which aligns with previous survey results. More specific questions in this category included: ‘How to write prompts to get the best responses from generative AI.’, ‘How to use generative AI for data analysis and visualization.’, ‘How to create visual materials like graphs and tables with generative AI.’.

Additionally, students expressed interest in knowing specific examples of how generative AI can be used in learning and research scenarios. Interestingly, many were curious about how their professors use generative AI. The second most common category of questions was about the "Permissible use of generative AI (29.7%)". Students wanted to know if using generative AI for assignments and exams is allowed, and if so, to what extent. This suggests that professors need to provide clear guidelines on the use of generative AI at the beginning of the semester or include such information in the syllabus. The third category was "Ethics of using generative AI(15.8%)“. Students wanted to know how much they can trust the outputs from generative AI, and they were interested in learning about the proper citation methods to avoid plagiarism when using generative AI. They were also interested in understanding how to develop critical thinking and autonomy without relying heavily on AI. Other notable categories of questions included: ‘Principles and information about generative AI (7.4%)’, ‘Fact-checking methods(5.3%)’, "Professors' perspectives on generative AI (3.4%)’.

  In this Teaching Tip, we analyzed the responses from surveys conducted in the second semester of 2023 and the first semester of 2024 to understand how students wish to learn about generative AI. The results show that students consistently expect guidance on "how to use generative AI for learning." They prefer to learn through its application in regular courses and are most curious about practical examples and specific methods of using generative AI. Students also expressed a strong desire for detailed guidance from professors on the permissible use of generative AI for assignments and exams. 

  The previous Teaching Tip indicated that as generative AI technology advances, students’ usage and satisfaction are also increasing. This teaching tip highlights the necessity for professors to actively guide and incorporate generative AI into their classes. Given its potential as a powerful tool in education, sharing practical examples and case studies on the use of generative AI in the classroom is crucial. In the next Teaching Tip, we will investigate and analyze examples of using generative AI in classes, and based on these findings, we will introduce ways for students to effectively utilize generative AI in their coursework.

Sang-eun Lee, Min-young Ku, Ye-jin Kim (2024), How do students want to learn about generative AI? (CTL Teaching Tips #43). Seoul: Sungkyunkwan University, Center for Teaching and Learning Innovation.