Can "Guided AI Use" Motivate Students to Learn Biology More Deeply?

My experience with 45 students' use of BIO181 GPT for inquiry-based learning

Preamble

In Fall 2024, I embarked on a pedagogical journey to innovate “BIO 181: Introduction to Biology and Biotechnology” using generative AI tools in collaboration with Dr. Emily Atieh (education researcher), Dr. Wei Li (learning technologist), and Will Hazelton (undergraduate teaching assistant). This essay reflects on my experience with 45 students, focusing on how students used and responded to the guided use of BIO181 GPT for inquiry-based learning of high-level concepts, in connection to real-world challenges. Here I would like to share newly gained insight.

Pedagogical Challenge

BIO 181 is a first-semester biology course which is offered to all majors and taken by over 400 students every year at Stevens Institute of Technology in Hoboken, New Jersey. Learning modern biology can be quite challenging for many students due to significant memory barriers to overcome before being able to visualize, comprehend, and integrate high-level concepts in real-life context. Most students quickly lose interest in learning biology deeply, despite its relevance and importance in our daily lives.

BIO181 GPT for Deeper Learning

I created BIO181 GPT by uploading relevant chapters of an open resource textbook (OpenStax Biology 2e) into ChatGPT 4o. The chapter-specific GPTs reside in ChatGPT Store, accessible through a learning management platform (Canvas), and were integrated into various learning activities throughout the semester. Here are several examples of these activities.

• Major Characteristics of Life: Students analyzed GPT response to determine if all major life characteristics were captured through an activity that compares a toad vs. concrete.

• Testable Hypotheses and Study Design: Students generated GPT-assisted hypotheses related to human health and discussed a GPT-generated scientific study designed to test a hypothesis of students’ interest.

• DNA to RNA to Proteins: Students used GPT to inquire about DNA’s molecular structure, replication, transcription to RNA, translation to proteins, and mutations that lead to genetic disorders such as cystic fibrosis.

• Gene Expression and Biotechnology: Students discussed the GPT-generated design of a scientific study relating to gene edited chickens for avian flu resistance from technical, ethical, and societal perspectives.

Student Feedback via Anonymous Survey

Student feedback was collected via an anonymous survey conducted toward the end of the semester. Survey questions were formulated to assess students’ motivation to learn in the context of the ARCS model of motivation – Attention, Relevance, Confidence, and Satisfaction (Keller, 1987). 42 out 45 students enrolled in the class responded to the survey.

• Attention: Just-in-time GPT inquiries appeared to stimulate inquiry arousal, with 72% of students reporting increased curiosity during activities.

• Relevance: 78% of students reported that GPT responses helped them relate biological concepts to real-life contexts, particularly for activities related to biotechnology (e.g., gene edited chicken for avian flu resistance).

• Confidence: 89% of students expressed increased confidence in completing active learning activities.

• Satisfaction: 67% of students reported that using GPT made learning more enjoyable, indicating the satisfaction of addressing problems independently with AI assistance.

Overall, the survey data suggest that the guided use of generative AI tools, like BIO181 GPT, can be used to enhance student motivation across all dimensions of the ARCS framework.

Areas for Improvement in Using GPTs

Students reported the following issues with

• Limited Prompting Skill: 20 students (44%) indicated difficulties with effective interactions with GPT.

• Information Overload: 15 students (33%) reported feeling overwhelmed by the volume of information provided by GPT, necessitating additional time to refine their queries.

• Accuracy Concern: 10 students (22%) encountered instances of inaccurate information provided by GPT.

Future Directions

I have become more optimistic that the guided use of generative AI in the classroom can help motivate students to learn biology more deeply. At the same time, I am concerned about the detrimental use of generative AI such as doing homework without reading the textbook. In Spring 2025, I am working on

• Provide more specific classroom instructions on the transparent and safe use of GPT.

• Tailor activities to address the clarity and accuracy of GPT response.

• Plan quantitative studies to assess the impact of AI-assisted learning on student performance and motivation.

I cannot wait to share our experience with my class in Spring 2025. Stay tuned!