What College Students Think About Artificial Intelligence

Source: Wes Hicks / Unsplash

Last month, I attended the DREAM conference in Philadelphia, and the buzziest conversations were around artificial intelligence (AI). Two keynotes and no fewer than a half dozen sessions discussed how AI is rapidly changing how students learn and complete assignments, and how colleges enroll, educate, support, and retain those students. Surely, no part of higher education will go untouched by AI.

One voice that was largely absent from these conversations was that of the students. Since the launch of GPT-4 in March 2023, there has been a pervasive fear among educators that students are more than eager to let AI do all their work for them and that we must be their shepherds of honesty and virtue. But is that true? Recent studies paint a far more nuanced picture of how college students think about and use AI.

Students think AI is great for gathering information

According to 2024 data, 86 percent of students already use AI in their studies, with 54 percent using it at least weekly. Moreover, 56 percent of students believe AI helps them learn, and 48 percent believe AI boosts their grades. But how do students use AI? Students appear to use AI for two distinct academic reasons: gathering information and generating responses.

The main reason why students use AI is to prepare for assignments and exams by gathering information/sources, summarizing notes, and validating facts. This use of AI is generally seen as acceptable. For example, students and teachers alike largely rated AI as OK to use during brainstorming and outlining as long as the use of AI is disclosed. In fact, Dr. Ethan Mollick of the Wharton School demonstrated at DREAM how he encourages his business students to use AI to generate entrepreneurial ideas for class projects.

Students think AI should not be used to generate responses

It should come as a relief that most students are not using AI to generate responses, including completing assignments, writing or rewriting papers, answering exam questions, or generating code. Most students see these uses of AI as cheating, especially when not disclosed. Students generally want to use AI ethically, with 80 percent saying that using AI to cheat should be punished as harshly as any other form of academic misconduct.

Another reason why students don’t use AI to generate responses is mistrust. Students in one study ranked the accuracy of ChatGPT as only 3.2 out of 5, and more than 70 percent said they would discontinue using AI if they learned it was less accurate than believed. Moreover, 43 percent of college-bound students said that using AI impairs critical thinking skills and creativity. So, students appear to be scrutinizing the information they gather from AI and are keen to do assignments for themselves.

This behavior could shift, however, as many honest students report worrying that their peers are gaining an advantage as early as the application process. These fears could foster pluralistic ignorance in which students erroneously believe that everyone else is using AI to cheat, therefore pressuring themselves into cheating to keep up. Educating students on AI, not banning AI, is key to preventing misconduct.

Questions to ask about AI in the college curriculum

• Is AI part of students' college skills? Nearly three-quarters of students want to better understand how AI works and how it can help them learn. AI literacy should be a mandatory topic within any college skills or first-year experience course. Creating open dialogues that norm how most students believe many uses of AI are academically dishonest will prevent pluralistic ignorance. Syllabi that set clear expectations for AI use by both students and teachers will be appreciated by students seeking guidance and assurance.
• Is AI part of faculty training? Nearly three-quarters of students also want their instructors trained in AI tools, yet more than 95 percent of teachers in one survey said their institution has provided no AI training whatsoever. Moreover, students generally approve of professors using AI to prepare course materials (e.g., creating images, building interactive content, outlining assignments) but not to create assignments and assessments, provide feedback, or assign grades. And same as they expect for themselves, students want faculty to disclose their use of AI. A shared reading experience or a professional development workshop on AI would be great places for faculty to start.
• Do you know what employers expect interns and graduates to know about AI? Colleges must align their educational standards with workforce expectations around AI. More than half of students want colleges to use more AI in teaching and learning, partly driven by the expectation that AI will be prominent in their future workplace. We must stay current on what employers want future employees to know about AI, and which skills are, in fact, being rendered obsolete by AI. These are important conversations to have even in fields in which the application of AI might not seem obvious.