Quick and Effective Ways to Design ChatGPT-proof Assignments

Tools like ChatGPT create new challenges that go beyond just integrating tech. Instead, it’s about keeping students engaged with assignments they can’t just “AI” their way through. Having spoken to plenty of teachers looking for new ways to teach in a post-ChatGPT world, we’ve collected various tips on how to “AI-proof” your assignments. 

Here’s a practical breakdown of how to adapt assessments to encourage genuine learning and minimize student reliance on AI.  

Strategies for AI-resistant assignments

To make assignments that AI tools can’t handle, you need creative prompts that ask for students’ unique input. Instead of generic questions, it’s about asking for answers that encourage students to justify their opinions and make connections across multiple sources. 

Get specific on the structure 

Instead of assigning a five-paragraph essay, set detailed guidelines that AI struggles to follow. For example: “In the second paragraph, the third sentence must introduce a counterargument,” or, “The final sentence in paragraph four should use terms from the topic sentence.” AI tools simply aren’t built to catch these types of specifics. 

Focus on timely content 

Personalized and real-world prompts make it a lot harder for AI to produce relevant answers. If students are reading The Hunger Games, ask for a comparison with a recent movie, like Black Panther – exploring themes like courage and challenging oppressive systems. This type of prompt asks for connections between the texts, getting students to dig between the lines for character motivation and societal messages (which AI struggles to do). 

Include Reflective components 

When students are asked to outline their thought process, it can deepen their understanding of a topic. Bringing in an element of reflection to an assignment can make AI assistance less feasible. For example, ask: “What challenges did you come across during this project, and what did you learn by overcoming them?”

Encourage process documentation

Ultimately, a core part of creating AI-proof assignments is encouraging the demonstration of a documented process. Get students to submit drafts, work versions, and annotated sources – this helps to build their research skills, and also shows exactly how they interacted with their sources. 

Lean on real-time, in-person engagement

Keeping assessments student-driven means testing their real-time thinking, understanding, and unique perspectives – by taking a creative and strategic approach. 

Use oral exams and presentations 

Oral exams test a student’s understanding and critical thinking in real time, in a way that AI cannot replicate. Making it a habit of having a brief Q&A after a presentation helps teachers to quickly gauge whether students have actually synthesized the information they have presented, or if they are simply reciting what an AI tool might have generated. 

Introduce open-book, open-note assessments 

Make sure assessments focus on that same element (synthesizing) and applying the knowledge – as opposed to simple rote memorization. An open-book assessment might ask, “Compare two historical leaders’ approaches to resolving conflict and discuss how these could be adapted today.” This type of format makes it much harder to use AI for an answer.  

Encourage peer review

Peer review means that students are engaging critically with one another’s work, and this element of having to discuss and defend their ideas in small groups can exercise their thinking in a way that AI cannot. In working alongside peers in this way, AI-generated materials also become simpler to spot.  

Test your prompts with AI tools

Before you set an assignment, run it through an AI tool like ChatGPT. If it is really easy for ChatGPT to generate a comprehensive response, keep altering the question until a deeper level of analysis or personal insight is required. 

Assignment types that go beyond AI’s reach

These assignments specifically tap into what AI doesn’t have: creativity, critical thinking, and personal engagement. 

Pen-and-paper assessments 

Handwritten essays that are completed in class show a student’s thought process as well as their authentic voice (which any AI tool would struggle to imitate). This can work especially well when it comes to foundational knowledge checks and timed essays.

Community projects

Projects that demand students do real-world exploration through surveys or interviews or local research mean that they connect directly with their local community. For example: “Explore the impact of a local environmental issue through surveys and firsthand interviews, and share your findings in a presentation.”

Personalized research 

This type of assignment asks students to conduct original research, collect data, or interview local experts – requiring them to then share personal insights and analysis based on that data. For example: “Choose a historical event or figure from the local community. Analyze their impact on the community and share your findings in a written report.”

Interactive debates or mock trials

Creating assignments where students participate in debates or take on roles in historical reenactments means that they have to think spontaneously. For example, “Organize a mock trial where students are in teams to represent the prosecution, defense, or jury.” They analyze evidence, develop persuasive arguments, and cross-examine witnesses while adapting to unexpected turns in the trial.

Flipping the classroom

Many teachers are already applying this approach, which is essentially about assigning reading material for homework, then getting students to complete assignments inside the classroom, where they can ask for guidance. This gives you more chances to support students directly. It also helps you give students opportunities to build their time management skills, as well as prioritization, selecting quality sources, and other tools they will need in the future. 

Use homework as preparation (not output) 

In a flipped classroom, homework isn’t the end product: it’s the preparation. Instead of assignments to complete at home, provide reading materials or background information that students need to look at before class. Then, when they come to class, they should be ready to engage and apply what they’ve learned, then and there. 

Class time for application and discussion

In class, students can work on tasks like group discussions, or problem-solving activities that require direct input and critical thinking. This means you can step in and give them guidance as required, and also primes the students to dive into more meaningful work while they’re in the classroom. 

Real-time assessment  

In a flipped classroom, assessments happen in real time. This makes it infinitely easier to gauge if there’s AI-generated or inauthentic work. In-class discussions, quick quizzes, and group activities can help you observe students’ comprehension levels and engage with them on the material, on a personal level.

Being open to AI 

Some teachers are embracing the rise of AI as part of their assignments, demonstrating through example to students how to use it responsibly. As our guest blogger Kellie explains, “Just like students quickly learn that Wikipedia is not a reliable, stand-alone source, they also need to know how to properly use ChatGPT and other content-generating tools and what their limitations are.”

For example, teachers might ask students to input a prompt into ChatGPT; include both the prompt and AI’s response in their work; then build on it with their own analysis or research. This demystifies AI and turns it into a learning tool rather than one that can provide a shortcut. It helps students gain valuable skills in evaluating information quality. 

A new era of AI-Resistant Learning

Although it is still an adjustment for many – AI is here to stay and is likely to play an increasingly vital role in how students engage in the world around them. This inevitably means that the way we are used to setting students up for success has shifted. While it is still early days for AI-resistant learning, educators are already finding creative ways to adapt using the strategies above. 

In a nutshell, embracing creativity and problem-solving helps students prepare for a world where AI is simply a tool rather than the answer itself. The above strategies also encourage students to integrate technology with their learning – without sacrificing the human aspects of education.