�Ƽ��Ӱ�ԭ��

In the dynamic landscape of artificial intelligence (AI) and education, integrating innovative technologies into the classroom is no longer optional—it is a necessity. As an AI enthusiast, I was thrilled when OpenAI launched the ability to build Custom GPTs in November 2023. I recall the excitement of staying up late that first night, much to the bemusement of my wife, as she wondered if I was ever going to stop “playing with that thing." Among the many ideas swirling around in my head about how to use this new feature, one stood out: personalized education. With that, my GPT teaching assistant named Wolfie was born. Wolfie was designed to provide individualized tutoring and examples tailored to each student's interests and would be available 24/7. For instance, if a student was into sports, Wolfie could give sports examples or analogies when explaining a value proposition; if a student was passionate about music, Wolfie could tailor explanations to include musical concepts when discussing the importance of branding and audience engagement in entrepreneurship. By leveraging this new technology, Wolfie ensures that each student's learning experience is more personalized and effective, aligning with the trend of adaptive learning.

Wolfie in the classroom: assignments and applications

Unfortunately, I wasn’t teaching in Spring 2024, so I had to wait until Summer 2024 before I could try out Wolfie with my Fundamentals of Entrepreneurship course. Given the asynchronous format of the three-week mini-term, I was particularly excited about the prospect of having Wolfie available 24/7 for students needing help with course materials. Before diving into the course content, I needed to understand my students’ baseline familiarity with AI, so I administered an AI Comfort Level survey. The results were a bit surprising — 70% of the 38 students reported having limited knowledge or experience with AI. To address this, I knew I would need to build assignments that not only taught entrepreneurship but also helped students become more comfortable with AI tools.

• Weekly Chapter Reflections: One of the first assignments where I integrated Wolfie was the weekly chapter reflections. Traditionally, these reflections would be posted on the Canvas discussion boards, followed by peer responses. However, I wanted to push students to engage more deeply with the material, so I added a requirement to engage in at least four prompts of dialogue with Wolfie. By engaging in a four-prompt dialogue with Wolfie, students were encouraged to think more critically and refine their thoughts. Over 85% of students reported that these interactions increased their engagement with the course material. Many noted that Wolfie’s feedback often challenged their initial thoughts, leading to more nuanced reflections.
  
  This is also one of the benefits of using a custom GPT.  I could train Wolfie to be more critical of their reflections and push them to move past surface-level responses. 

• Shark Tank Elevator Pitch: During the first week of the course, OpenAI launched GPT-4o, which introduced —a new feature that enabled more natural, conversational interactions with the AI. This upgrade couldn’t have come at a better time, as the students had a Shark Tank-style Pitch assignment due in week three. Initially, this assignment required students to record a simple 1-1:30 minute pitch. However, with the introduction of Voice Mode, I saw an opportunity to elevate the exercise to a new level of engagement and realism.
  
  Instead of just recording their pitch, I had students engage in a dynamic negotiation with Wolfie, who adopted the persona of seasoned Shark Tank investors. Embodying the assertiveness of Kevin O’Leary or Mark Cuban, Wolfie challenged them to defend their ideas and forced them to negotiate equity stakes in real-time. The probing questions pushed them to think critically, as one student noted, “For the elevator pitch, the AI helped a lot because it asks good questions I never thought of asking.  It really forced me to think more about my idea in a way I hadn’t at the beginning.” Overall, 80% of students felt that this exercise improved their pitching skills, demonstrating the potential of AI to enrich traditional assignments.

• Final Business Plan Review: As we approached the end of the course, I introduced a new element to the final business plan assignment—AI-driven feedback. I asked each group to submit their rough draft to me and then to Wolfie for feedback before turning in the final version. Many students were initially unaware that they could prompt Wolfie to act as a writing tutor or review their plan and give them a grade based on its interpretation of my rubrics. This again highlights the unique strength of a custom GPT, as I added my rubrics to Wolfie’s knowledge base and also instructed it to be particularly critical, not giving any grade higher than 90% in its feedback. My goal with this assignment was to be able to provide immediate, detailed, real-time feedback that forced students to think more critically about their ideas.
  
  The editing and reviewer capabilities were particularly praised in the end-of-class survey, with 85% of students indicating they appreciated Wolfie’s thorough feedback and felt it significantly improved their final submissions. As one student put it, "The AI's feedback was more detailed than what I usually receive in peer reviews, and it helped us refine our business plan significantly."

Reflections and future applications

Integrating Wolfie into my summer class was both an exciting and intriguing experience. Despite some technical and ethical challenges, 88% of students were satisfied with their experience using Wolfie and would recommend it for future courses. This positive outcome has inspired me to push the boundaries further in Fall 2024, where students in Bus 101 will take the lead by creating their own AI assistants and tutors.                                                                                                       

However, it’s crucial to approach this next step with care, ensuring that AI serves to enhance—not replace—the human elements of teaching and learning. Striking the right balance between technological innovation and the irreplaceable value of human interaction will be essential to fully realizing AI's potential in education.

�Ƽ��Ӱ�ԭ�� the author

Jeremy Tiedt is the assistant dean of undergraduate student success in the College of Business, where he also teaches undergraduate business courses. He is pursuing a doctorate in information technology within the College of Education, focusing on the integration of AI in educational technology.