University Honors

Artificial intelligence (AI) use is becoming increasingly popular among students in higher education. Many students utilize AI to assist with tasks such as brainstorming, drafting, editing, summarizing information, studying, and note-taking. Of particular interest is understanding the ways that AI use can help or hinder note-taking. Previous research has shown that note-taking is beneficial to learning because it requires people to put information into their own words, which engages cognitive processes that allow them to gain a deeper understanding of the materials. Using AI can help students with summarizing the main ideas of the learning materials, which supports their learning process. Additionally, it may reduce their test-taking stress by making them feel more prepared for the exam. However, AI use may also hinder learning if it replaces the cognitive processes involved in note-taking. Therefore, this study addressed: 1) whether using AI to assist with note-taking is more effective at supporting learning than traditional note-taking, and 2) whether having support from AI during learning can alleviate potential stress and help students feel more prepared for an exam. In this study, participants read a passage and took a knowledge test under one of three conditions: control (read-only), note-taking, and AI-assisted note-taking (using AI to identify the main points of the passage in conjunction with their own notes). Then, participants answered questions about their levels of stress and preparedness given their test preparation method. I hypothesized that the note-taking group would have the highest memory performance because note-taking requires students to paraphrase the information, which supports memory retention. However, AIís ability to help students strengthen their notes could potentially allow the AI-assisted group to improve their performance relative to traditional note-taking. Additionally, I hypothesized that the AI-assisted group would experience less stress and feel more prepared than the other two groups, and thus, have better test performance due to higher confidence levels in their knowledge of the content. However, using AI could also be more stressful since the information generated by AI might not be detailed enough to help students do well on the test. Contrary to my hypotheses, there was no significant difference in test scores or stress levels between the groups. However, the AI-assisted group reported feeling significantly more prepared for the exam than the read-only group. Though it is still unclear whether AI use is more beneficial or harmful to learning, this study was a critical first step in determining how AI use could potentially benefit students in higher education.