My GPT for Grading and Feedback

From Theory to Application

A friend invited me to a workshop at Fordham University designed to help teachers understand, adapt, and embrace the use of AI in the classroom. Bravo to Fordham for for the foresight and wistom to sponsor this necessary workshop and thank you for allowing me to join the workshop.

When the facilitator, Tal Goldhamer, demonstrated how to create GPTs and provided several examples of their potential applications, such as grading papers, I perked up. While I have lots of experience with Large-scale GPTs like ChatGPT, Gemini, Copilot, Claude, and Perplexity can can generate text, translate languages, create diverse creative content, provide informative answers, and even engage in conversational interactions, I had yet to dabble in creating my own.

Creating a Custom GPT for Grading

One of the most surprising aspects of this experience was the realization that creating a custom GPT is within reach for educators like myself. The technical process is remarkably straightforward – you add instructions (prompts) and hit create. However, the true challenge lies in the crafting of prompts and the continuous refinement of the system. It's a testament to the democratization of AI technology, opening up new possibilities for educators willing to explore.

Aligning AI with Classroom Expectations

Even thought the creation of the GPT app is technically simple, the devil is in the details and is only as good as my prompts, here are some highlights of my prompts, which come from my rubric, assignment and other details shared with my students.

• Emphasis on Voice: I prioritize the development of students' unique perspectives and ensure their work is both accurate and engaging.

• Clear Expectations: By breaking down elements like Understanding, Organization, and Polish, I train my students (and the AI) and set clear expectations for depth and quality in their analysis.

• Alignment with Goals: This structure not only makes it easier for students to understand how to succeed but also aligns well with my goal of nurturing thought leaders, not just tacticians.

• Expectation for Reflection: When grading your students' assignments, I expect their work to directly reflect what we've discussed and covered in class.

• Avoidance of Generic Responses: I'm not looking for a regurgitation of general internet-based definitions or common ideas about digital marketing.

• Critical Thinking: The goal is to ensure students think critically within the scope of our class discussions and produce work that's in line with what we've learned together.

After I typed in my prompt, I hit 'Create.' Poof! The GPT appeared like magic! To grade, I just dropped a student's PDF on the GPT's desk and said, 'Do your thing, grading bot!' 🤖📚

How it works

The GPT works by first checking the assignments against the rubric. It then gives students scores and feedback to improve based on what's in the rubric. I review everything the GPT does and make any necessary changes to make sure the feedback is fair. So if I saw something that was above and beyond my expectations for the students of what I had taught, I tell the GPT that it is covering something I didn’t teach yet and it adapts. As more assignments are graded, it gets better and better. To ensure that it is as fair as possible, I did this process for two rounds for each class.

The Human Touch in AI Implementation

While the creation of a GPT might be simple, using it effectively in education requires careful consideration and constant oversight. My experience highlighted the critical importance of human supervision in AI-assisted grading. The GPT is capable of providing consistent evaluations, but it was my role to ensure that the feedback aligned with the nuances of student work and the broader educational goals.

A Teaching Moment

One of the most exciting outcomes of this experiment is its potential as a learning tool. By openly discussing the process with my students and sharing this article as a model of the digital perspectives I expect them to develop, I've transformed it into a hands-on lesson in digital marketing, curiosity, and innovations. It showcases the practical application of AI in marketing, providing students with direct experience of the technology that will define their future careers.

Ethical Considerations and Student Feedback

Implementing AI in grading raises important ethical questions about fairness, bias, and the role of human judgment in education. To address these concerns, in class I will make it clear to my students that this system is part of an ongoing experiment. They will have the opportunity to review the process in class and provide feedback, especially if they feel any aspect of the grading was unfair. This openness not only maintains trust but also engages students in critical discussions about AI ethics.

Enhanced Feedback and Time Management

One unexpected benefit of this system was the ability to provide more comprehensive feedback. The GPT-assisted grading offered not only specific grades on various criteria but also suggestions for improvement and positive reinforcement – often more than I would typically manage on my own. This freed up time for me to engage in higher-level educational activities, like writing this article to share my experiences and insights.

Looking Forward

Our goal is not just to prepare students for the jobs of tomorrow but to empower them to bridge the gap between visionary possibilities and current business realities. In doing so, we're cultivating a generation of digital marketers who can navigate the complexities of our field with both innovation and practicality, ready to lead the digital transformation that many companies so desperately need.