Introduction

While ChatGPT can be a valuable learning tool, it has also made cheating easier for students. To combat this, educators can develop strategies to “ChatGPT-proof” their courses. In this article, we will discuss best practices for minimizing academic dishonesty related to ChatGPT and other AI language models.

This chapter is divided in the following sections. Click the links below to jump down to a section that interests you:

• Encourage Critical Thinking and Originality
• Implement Student Collaboration and Group Work
• Utilize In-class Assignments and Exams
• Use Plagiarism Detection Software
• Foster a Culture of Academic Integrity
• Lean Into the Creative Use of ChatGPT
• Conclusion

Encourage Critical Thinking and Originality

One way to deter students from using ChatGPT to complete assignments is to design tasks that emphasize critical thinking, problem-solving, and originality. Here are a few suggestions:

Promoting Critical Thinking through Open-Ended Questions

Assign open-ended questions that require students to form their own opinions, analyze multiple perspectives, and provide evidence to support their viewpoints.

Applying Concepts through Real-World Scenarios

Incorporate case studies or real-world scenarios that challenge students to apply concepts to new situations.

Fostering Collaboration through Project-Based Learning

Use project-based learning, where students must collaborate to create unique solutions to complex problems.

Implement Student Collaboration and Group Work

Assignments that involve collaboration and group work can help discourage students from relying on ChatGPT. By working together, students can learn from one another and develop a deeper understanding of the course material. Here are some ideas for group-based assignments:

Collaborative Learning through Group Presentations

Assign group presentations, where students must collaboratively research, analyze, and present on a topic.

Developing Products through Group Projects

Create group projects that require students to develop a product, such as a research paper, website, or prototype, based on their collective knowledge and skills.

Organize peer-review sessions, where students provide feedback on one another’s work, promoting a culture of academic integrity.

Utilize In-class Assignments and Exams

In-class assignments and exams can reduce the likelihood of students using ChatGPT to complete their work. Consider these strategies:

• Implement short in-class writing assignments, where students must respond to prompts and provide thoughtful analysis.
• Use timed exams or quizzes that make it difficult for students to rely on AI-generated assistance.
• Include application-based questions that require students to analyze and interpret specific examples from course material.

Use Plagiarism Detection Software

While ChatGPT-generated text may not always be detected by traditional plagiarism checkers, these tools can still help identify instances of academic dishonesty. In addition to using plagiarism detection software:

• Encourage students to submit drafts of their work for review, making it more difficult for them to rely on ChatGPT.
• Provide clear guidelines on citation practices and emphasize the importance of academic integrity.

Foster a Culture of Academic Integrity

Promoting a culture of academic integrity can help prevent students from seeking AI-generated assistance. To do this:

• Clearly communicate the expectations and consequences related to academic dishonesty, including the use of AI-generated content.
• Encourage open dialogue about the ethical implications of using AI tools like ChatGPT for cheating.
• Integrate discussions about academic integrity into your course and emphasize its importance in the broader context of higher education and professional life.

Lean Into the Creative Use of ChatGPT

Instead of preventing students from using ChatGPT, consider structuring assignments that require them to use the platform in ways consistent with their future profession.

For instance, I teach an Introduction to Educational Technology course, primarily for undergraduate education majors aspiring to become teachers. They must complete various assignments that involve designing lesson plans for any grade level and subject area, integrating suitable educational technology tools.

I aimed to create several assignments so that students learn to utilize ChatGPT effectively in their future teaching careers.

Here are a few examples of assignments I developed:

Assignment Instructions #1

For this activity, you will revisit the Lab 6 activity from several weeks ago where you had to craft a lesson plan that used some form of blended learning. This time around, however, I would like you to try using ChatGPT (or some other Chat AI tool) in order to generate and refine a lesson plan.

Using a Chat AI tool, develop a prompt that will generate a usable lesson plan in your grade level and subject area on a topic of your choosing that integrates some form of blended learning (like you created in the initial assignment from week 6).

After GPT generates a response, read it over and consider the following questions: Where are the weak spots in this lesson plan that it created? What additional resources would you need in order to actually teach this lesson with students? Ask ChatGPT a series of follow up questions in order to refine/revise the output and the resources that you have available to actually teach this lesson.

Your final submission should include all of the back and forth conversation that you had with ChatGPT to generate this outcome. You can simply cut and paste the responses into a Word document like you see in the attachment below. You should include at least four follow up questions after your initial prompt along with ChatGPT’s responses.

At the end of the submission, include your own reflection (not generated by ChatGPT) where you consider the following:

• What are the major benefits and challenges that this Chat AI technology provides to students?
• What are the major benefits and challenges that this Chat AI technology provides to teachers?

Assignment Instructions #2

For this activity, perform the following steps:

1. Decide on a function related to teaching and training that you would like the LLM to perform. This can be related to the role of an educator (i.e., learning designer, teacher, etc) or the role of a learner. For example, you might want the LLM to help you to do one of the following tasks:
   1. Educator Role: Develop a lesson plan for a grade level and content area of your choosing.
   2. Educator Role: Prepare a presentation on a topic that you could be asked to deliver.
   3. Educator Role: Create a series of learning objectives based on the title of a course.
   4. Educator Role: Develop an activity or assessment for a specific lesson you might teach.
   5. Learner Role: Provide feedback on a piece of writing you drafted.
   6. Learner Role: Ask for assistance with studying for topic you are interested in learning.
   7. Learner Role: Help you to learn a foreign language by conversing with you and providing corrective feedback.
2. Develop a prompt that you will provide the LLM in order to execute the task. Click on the expander below to open it up and read a note about generating better content in an LLM.
3. Feed the prompt you developed into the LLM. You might try out more than one prompt with varying degrees of specificity in order to examine how they differ and which ones you like better.
4. Evaluate the results it produces in order to see what you like and what you don’t like. This step does not need to be documented in your final submission.
5. Ask a series of follow-up requests that attempt to refine and improve on the output. For example, if you asked to generate a student activity, you might ask it to create a rubric that could be used to evaluate the assessment. You might ask it to revise certain parts or develop other parts in greater detail. Some of the best output that LLMs like ChatGPT produce occur in these follow up requests that help it to refine the final product in a manner that matches precisely what you are looking for. You should include a minimum of three follow up requests that occur after your initial prompt.
6. Copy and paste all of your prompts, follow-up questions, and the output that is generated into the template included below.
7. Finally, write up a personal reflection about how the process worked for you. This should not be generated by the LLM but reflect your own thoughts and experiences. You might reflect on some (or all) of the following questions in your response:
   1. Was the output better or worse than what you expected?
   2. What surprised you?
   3. How might you see yourself integrating this tool in your work as an educator?
   4. How does this experience affect your perception about the extent to which AI and LLMs are either a net positive or net negative for the field of teaching and training?

Conclusion

To minimize the potential for cheating with ChatGPT and other AI language models, faculty members in higher education must adapt their course designs and assignments. By promoting critical thinking, originality, collaboration, and academic integrity, educators can create a learning environment where students are less likely to rely on AI-generated assistance. Ultimately, these strategies can help ensure that students develop the skills and knowledge necessary for success in their academic and professional lives.