Abstract 2122 Use of ChatGPT to generate study materials in an introductory biochemistry course

While taking notes is generally accepted as a good skill for students to develop and the act of taking notes in and of itself assists in learning, it can be a barrier to success for some students. Over the past year Large Language Model (LLM) artificial intelligence programs such as ChatGPT have become widely available. These programs are effective and efficient at performing natural language processing (NLP) tasks. Examples of these include many of the tasks commonly discussed in the media, including generation of text, translation, and AI based assistants. This study sought to determine how effective ChatGPT 3.5 was in generating study materials for an introductory biochemistry I course. A short lecture on beta oxidation was delivered in class. The lecture was transcribed using MS word. The resulting transcript was curated to remove paralanguage, then submitted to ChatGPT 3.5 to generate a summary, a set of notes, and an outline. These were each examined for accuracy by the lecturer. ChatGPT was easily able to generate these documents in a very short time frame (seconds to minutes). Each of these four artifacts were provided to students to help them prepare for a quiz. Students were asked about the use of AI generated materials compared to other study materials (e.g., lecture notes, study guides, homework problems, and in class exercises). Student use of the AI generated materials was somewhat bimodal with some students failing to use them while others found them as useful as the text or their own class notes. In conclusion, ChatGPT rapidly generated accurate study materials from a transcript. This presents an interesting opportunity for the development of novel study tools.