Beating ChatGPT with ChatGPT: Using AI Technology to Create Authentic Assignments 

David Woodring, PhD, is a criminologist/medical sociologist who currently serves as an adjunct instructor for Southern New Hampshire University, Eastern Gateway Community College, and Northwest Arkansas Community College, guiding students across a variety of subjects from cultural awareness in online learning to introductory sociology and social problems. Dr. Woodring also serves as a consultant for, where he teaches instructors/educators how to enhance teaching, learning, and instructional design with new digital technologies, including artificial intelligence. 

Author David Woodring smiling with his arms crossed.
David Woodring
Courtesy of David Woodring

We’ve all seen the headlines—students are using ChatGPT more than we know. Sometimes it can feel like we’ve lost the battle before it’s even begun, but there is a way to use ChatGPT to beat ChatGPT. I spent last semester using ChatGPT to help me create more authentic assignments that engaged my students, discouraged them from using ChatGPT to “cheat,” and led to a deeper understanding of the course material.  

In this blog post, I’ll share a few ways that you can use ChatGPT to design personalized and authentic assignments that encourage higher-order thinking and are inclusive and equity-minded. I’ll also share a few case studies that reflect my own teaching and offer tips for how to prompt ChatGPT in ways that align with your teaching goals.   

Before we begin, a few caveats: ChatGPT is a great teaching assistant. However, it can’t replace the knowledge and understanding that is required to create effective teaching and learning activities. While ChatGPT is trained to be inclusive and equitable in creating activities, biases can still emerge. Always review ChatGPT’s outputs and don’t rely on it to replace fundamental aspects of instructional design, teaching, or learning. In addition, while OpenAI has implemented privacy measures, it’s important to ensure that sensitive information is not input into the AI system.  

Now, with those caveats in mind, let’s discuss how to make the most of ChatGPT in class.   

Guiding ChatGPT with My Teaching Philosophy
To start my last semester, I gave ChatGPT the guiding principles from my teaching philosophy and asked it to help me create authentic and customized assignments for an introduction to sociology course based on my teaching preferences and approaches. Whether you emphasize constructivism, student-centered learning, or another pedagogical approach, ChatGPT can generate a wide variety of authentic activities. It may even introduce you to new philosophies, practices, and principles that enhance your skills as an educator.  

In this case, I needed an assignment that covered the weekly readings and that students couldn’t simply paste into ChatGPT for answers.    

Here is a modified version of the prompt I used: I need to create an authentic assignment that covers the sociological theories, terms, and concepts from the weekly readings in my intro sociology course. I need to design questions that students can’t simply paste into ChatGPT and generate answers. Let’s work on designing this assignment for the section on Constructing Gender/Sexuality. The assignment should be based on the following guiding principles from my teaching philosophy:  

  • Collaborative Intelligence
  • Real-World Problem-Solving
  • Active Knowledge-Making
  • Productive Diversity
  • Differentiated Instruction
  • Metacognition
  • Multimodality
  • Authentic & Continuous Assessment
  • Civic Pluralism & Competence

 After the iterative process of going back and forth with ChatGPT over different question types, it wasn’t long until we created Personalized Connection Questions (PCQs). For example:  

  • In what ways has the Me Too movement changed your own understanding of gender and power dynamics in our society? Has it impacted your personal relationships or experiences in any way? 
  • Have you ever witnessed or experienced the negative effects of hegemonic or toxic masculinity (such as aggression, violence, or objectification)? How did these experiences shape your understanding of masculinity and power dynamics in our society? 

This was great for concepts/terms from the course, but I also needed them to engage with theories. ChatGPT and I then created Personalized Theory Application Questions (PTAQs). For example:

  • Think of an example from the media where you’ve seen evidence of gender inequality (such as in a movie, TV show, or news article). How might feminist theory explain this example, and in what ways might that explanation overlook important aspects of gender inequality in this situation? 

I still needed to design the overall project around my teaching philosophy. After creating the questions over the key terms, ChatGPT was then able to give me several options for assignments that included my guiding principles and, without asking, created a rubric for me. We ended up calling it the Personalized, Engaging, Application, and Reflection (PEAR) assignment. The reflection questions were a key component that allowed formative assessment of key theories, terms, and concepts. I received positive feedback from students reflecting about their own learning and the assignment itself. For example, one student noted: “I love assignments that ask about my own life and experiences and how I can relate them to what I’m learning.”  

Beating ChatGPT with Authentic Assignments 
While I was happy with the personalized/authentic components of the PEAR assignment, I still needed to scaffold an authentic project that fostered higher-order thinking skills: analysis, evaluation, and creation. The project needed to have real-world relevance; be multi-step, collaborative, and open-ended; and have a peer-review/self-reflection component.  

Prompt: Help me brainstorm a project to scaffold on top of weekly readings, lectures, and the above assignments we created that uses higher-order thinking skills as outlined by Bloom’s Taxonomy, Webb’s Depth of Knowledge, and Costa’s Levels of Questioning. The assignment should be a multi-step, application-based, problem-solving project with real-world relevance. The project should give students real-world experience/skills that are sociological, general, and practical. It should start with the section on research methods and unfold throughout the semester with the typical logic/framework of an intro sociology course.  

While ChatGPT provided suggestions for a semester-long Sociological Snapshot assignment, I used the response to brainstorm and create a real-world career scenario where students could see how sociology could be relevant to their future careers, as discussed in the following section.     

Advancing Student Knowledge, Skills, and Abilities with ChatGPT 
Given my background in research, expertise in data analysis, and the fact that we were starting the section on research methods the following week, after brainstorming I asked ChatGPT if it could make sample data for a mock research study. Click here for the full prompt I used.   

Within just a short time I had partnered with ChatGPT to build an eight-week multi-step, problem-based, and application-based assignment that simulated conducting a real-world research study. The study followed the scientific method, and ChatGPT helped me create a simulation video that welcomed students to research positions at a state-level agency. I created a brief but realistic hiring/onboarding process where students submitted resumes and received/returned job acceptance letters.  

The most powerful part of the assignment was that ChatGPT created instant sample qualitative/quantitative data based on student topics/questions. ChatGPT provided data based on my parameters and was able to help me construct real-world models and scenarios that students would encounter if they were sociologists conducting research. ChatGPT was able to create rubrics for each stage of the assignment.   

Prompt: Create a sample dataset with 100 observations containing the following variables: ‘ID’ (unique identifier), ‘Age’ (numeric variable representing age in years), ‘Income’ (numeric variable representing annual income in dollars), and ‘Education’ (numeric variable representing years of education). The values for ‘Age’ should range from 20 to 60, ‘Income’ should range from 20,000 to 100,000, and ‘Education’ should range from 8 to 20. Ensure that the dataset has a mix of random values within these ranges for each variable. Format the dataset as a comma-separated values (CSV) file for easy import into Stata.  

Note: I asked ChatGPT to create qualitative data based on the students’ interview/focus group questions, and then I asked it to analyze the data for emergent themes using grounded analysis.   

I was amazed as I watched the most powerful teaching/learning activity I had ever created simulate the research process for students. Students worked individually and in teams to brainstorm, conduct research, give each other feedback, and then interpret the findings from the results of the data analysis I provided them. I allowed them to use ChatGPT during the research process at key points in the semester to show them the merits and drawbacks of ChatGPT and how to use it responsibly and ethically. At the end of eight weeks, I watched undergraduate students give academic presentations where they presented previous research findings, interpreted/applied theories, interpreted p-values, and presented their own research findings. Halfway through the assignment, students were engaging with sociology on a deeper level than I had witnessed in previous classes.    

And the students noticed a difference as well. As one student noted: “I didn’t realize this much work went into sociological research. I thought sociologists just sat around, observed, and took notes about people.”   

Designing Equity-Minded Assignments
As I guided ChatGPT with my teaching philosophy to create real-world authentic assignments that fostered higher-order thinking, I realized what a powerful tool ChatGPT was for creating personalized, equity-minded content. I was able to input nonidentifying information about my students from a student questionnaire I collected at the beginning of the semester to facilitate designing personalized content that was relevant to their interests and goals, and that was culturally appropriate and responsive.   

I generated multiple versions of various assignments, based on different learning styles, levels of understanding, and performance. This was especially helpful for students who missed active or collaborative learning components of class, because ChatGPT was able to quickly create alternative assignments that were constructively aligned.   

After a couple prompts asking about diversity and inclusion, ChatGPT offered multiple suggestions that helped to avoid biased language, considered accessibility, considered language notes, as well as personalized feedback based on student work. By the end of the semester my course was more inclusive and equitable.  

Prompt: Here are nonidentifying data on interests/backgrounds of the students in my course based on surveys/questionnaires. Please check this PEAR assignment to make sure it promotes fairness, inclusivity, and equal opportunities for all students. Offer any additional suggestions, insights, and/or recommendations for making the assignment more equitable, inclusive, diverse, and/or culturally appropriate/responsive.   

Leveraging ChatGPT to create authentic assignments has been a transformative experience, enabling personalized and equity-minded learning. By harnessing AI technology, I designed assignments that fostered higher-order thinking, engaged students, and promoted inclusivity and diversity. ChatGPT served as a valuable tool, generating content aligned with my teaching philosophy and helping students develop critical thinking skills. Moreover, by incorporating nonidentifying student data, ChatGPT facilitated the creation of culturally appropriate and inclusive assignments. Through this process, we can use technology to enhance teaching materials and practices and ensure equitable and authentic learning experiences. It is clear that ChatGPT can be part of a more holistic approach to teaching/learning that creates learning environments where students don’t rely on ChatGPT and creates assignments students can’t easily paste into ChatGPT for answers.   

One thought on “Beating ChatGPT with ChatGPT: Using AI Technology to Create Authentic Assignments 

  1. Thank you for this! I was giving you the side eye at first, since as an English professor who teaches writing, I have found the latest developments in language model AI like ChatGPT to be anxiety provoking. I would have not dreamed of befriending it to create learning opportunities in tandem. I am not sure whether I am brave enough to attempt a full partnership in assignment design this close to the first day of classes, but I will definitely be including much more personalization and collaboration in my writing assignments for students this term. Blessings on your work!

Leave a Reply