While working with a group on designing choice boards, a teacher expressed concern about maintaining the integrity of student work in an era of AI. She did not want students to rely on AI to produce the work for them. She worried that if students were self-pacing through a choice board while she pulled small groups for differentiated instruction or intervention, they might turn to AI tools to complete assignments rather than doing the original thinking themselves. Her question sparked a conversation about designing learning tasks that foster creativity, critical thinking, and problem-solving—skills AI still cannot do as well as humans.
I frequently hear variations of this concern about students creating authentic work when they can access AI. To address this concern, I want to review strategies for designing AI-resistant tasks and prompts that can be included in a choice board, self-paced playlist, or even a learning activity at a station in a rotation. It’s important to note that no task is entirely foolproof. Students may still find ways to bypass these strategies, so I avoid using the term “AI-proof.” Instead, these strategies are designed to raise the bar, making it more difficult for students to rely on AI alone and, instead, to encourage original thinking.
#1 Require Personal Reflections or Experiences
AI does not have personal experiences or emotions to draw from, so humans are better at personal reflection and authentic storytelling. Educators can capitalize on this shortcoming by asking students to respond to questions and tasks that require them to reflect and connect their learning to their lived experiences. Not only will this make it harder for an AI chatbot to complete their work, but it is also an opportunity to help students think more deeply about their learning and make meaningful connections between their classes and their lives.
Example: Instead of asking students to describe the entire cell division process, which would be an easy explanation to generate with AI. Science teachers could ask students to identify the step in cell division that was most challenging for them to learn and ask the students to describe the process they used to understand it or the analogy they created to make sense of this step in the process.
#2 Incorporate Creative Tasks
Encouraging students to create offline is an effective way to limit AI’s role in completing assignments. Tasks that require students to produce physical artifacts—such as building models, drawing visual representations, or creating artistic interpretations of their learning—help minimize the temptation to rely on AI tools.
Even if part of the task involves online work, like reading digital texts, watching videos, listening to audio recordings, or conducting research, teachers can pair that online work with an offline product to require students to actively engage with the material and surface their thinking and learning in creative ways.
Example: A social studies teacher asking students to research a particular moment in history or time period can ask them to surface their learning by creating an artistic timeline on a piece of paper that combines drawings, dates, and key information. Similarly, a history teacher asking students to listen to a podcast related to a topic they are learning about can encourage students to create a visual representation of what they are learning in the form of a concept map or sketchnote.
#3 Require Use of Specific Local, School, or Community Resources
Anchoring assignments in a local, school, or community context requires details and information that AI may not have access to. This can enhance student engagement as they think about the material they are learning through the lens of their lived experience. Asking students to analyze and reflect on resources unique to their environment encourages critical thinking and a stronger personal connection to the task. This approach can also lead to more authentic, meaningful learning experiences grounded in the real world.
Example: An English teacher can assign students to write an argumentative essay or prepare for a debate about a specific issue affecting their school or community (e.g., addressing a local environmental challenge or making the case to change an established school policy). Tasks and assignments that require students to gather information from school administrators, community leaders, or local organizations become more personalized and grounded in real-world concerns.
#4 Incorporate Peer Interaction or Collaboration
When students work together, they benefit from their peers’ support and diverse perspectives, reducing their reliance on AI to complete tasks. Collaborative activities—like group discussions, peer feedback sessions, or cooperative learning tasks—result in a learning environment where students share ideas and problem-solve. This process requires active engagement, communication, and social negotiation, which fosters deeper understanding and makes the learning experience more dynamic.
Students are less tempted to turn to AI for shortcuts when they can rely on their peers to clarify concepts, explore different solutions, and build on each other’s ideas. Collaboration also encourages accountability to their peers, making it harder for them to outsource their work to AI.
Example: In a world languages class, students can work in small groups to watch a video or analyze pieces of artwork from a culture related to the language they are studying. Using the reciprocal teaching strategy, each student would take on a role—summarizer, questioner, clarifier, or predictor—to guide the group’s discussion of the media in the target language. For example, the summarizer would summarize key ideas in the video or artwork in the target language, while the questioner asks questions about its cultural significance or themes.
#5 Challenge Students to Justify Their Process
Asking students to explain and justify their decision-making process encourages deeper reflection and critical thinking, making it much harder for AI to replicate their work. When students must articulate how they arrived at a solution, what strategies they used, and why they made certain choices, it pushes them to stretch their metacognitive muscles and think about their thinking in an intentional way.
Understanding our students’ process also allows us to assess not only the end product but the thinking behind it, providing invaluable insights into their learning. Since AI can generate answers but cannot effectively explain personal thought processes, this approach reduces reliance on external tools, leading to more authentic, thoughtful work.
Example: Instead of simply asking students to complete a set of math problems, a math teacher can challenge them to record a short video explaining how they solved a problem as if teaching it to a younger student. In the video, students would walk through the process step-by-step, using examples and language that are accessible and easy for younger students to understand.
Wrap Up
Designing AI-resistant tasks isn’t about avoiding technology; rather, it’s about designing assignments that challenge students to think critically, collaborate meaningfully, and demonstrate their unique understanding in ways AI simply can’t replicate. By encouraging offline creation, rooting tasks in local contexts, fostering peer collaboration, and asking students to justify their thinking, we position them to take ownership of their learning. The goal is to architect authentic learning experiences where students actively engage in problem-solving, creativity, and collaboration.
For more on creating AI-resistant writing prompts, check out my conversation with Dr. Randy Laist, Dr. Cynthia Murphy, and Dr. Nicole Brewer about their book Rethinking Writing Instruction in the Age of AI.
6 Responses
AI it seems has become an enemy to some, when in reality, we have to learn how to grow with technology as it continues to advance. I struggle with AI when students use it to do their work for them, but love using it as a quick guide to resources, as a tool to change reading levels for students, or as a way to increase my own knowledge in the classroom. Do you have any ways to incorporate AI appropriately into tasks versus trying to avoid it?
Hi Kylee,
I could not agree more! I know there is a lot of resistance to AI because teachers worry students may use it to generate work for them, but it has so much potential to support learners.
I wrote a blog about teaching in harmony with AI and in my book Shift Writing into the Classroom, I highlight an AI-enhanced strategy to support each step in the writing process.
You are making me think I should dedicate a blog (or series) to the reverse topic…how to create tasks that ask students to use AI appropriately for specific tasks.
Thank you for the comment and the idea!
Take care.
Catlin
Hi Catlin, I’m a master’s student in a secondary educator preparation program and I’ve been thinking a lot about how AI is influencing our practice as teachers. A week ago, I was trying to explain to some colleagues how I identify AI-generated writing and I came up with the following list:
-Large language models do not use figurative language unless prompted to do so. There will be a lack of jokes, idioms, metaphors, sarcasm, or pop culture references in writing that has been AI-generated. All of the writing will be very (overly) literal and direct. (For this reason, AI writing may be easier to detect in the humanities and harder to detect in STEM disciplines, which require more direct and literal writing!)
-In a similar vein, AI cannot give personal anecdotes or relate things to its own life because it is not alive and thus does not have life experience! AI writing is often impersonal and devoid of emotion.
-AI-generated writing tends to overuse words that indicate sequence or transition. The ideas in the writing may almost seem TOO organized sometimes– it can read as choppy and lack flow. (Bulleted responses are a dead giveaway for ChatGPT!)
-AI can only summarize or regurgitate information that it has received as input; it doesn’t come up with any new ideas (or if it is prompted to, it often results in bizarre or nonsensical output)!
Reading through this blog post, I was reminded of my list because both you and I were thinking about what AI cannot do that humans CAN do. While using technology as a tool to help with work is a normal part of 21st century life, using technology INSTEAD of doing work is something that will ultimately harm our students’ self-efficacy and intellectual development. Embracing an asset mindset means that we need to think of our students as irreplaceable by robots… which means we shouldn’t assign them robotic work!
Out of all of the recommendations you made, I think that asking students to reflect on and justify their work is the most practical method we can implement in our classrooms, since it can be done before students have started their work, during their work, or after they have completed their work. Novice teachers like me who have had less time and experience lesson planning and are more likely to be using inherited activities from other teachers may struggle to come up with new, AI-resistant lesson plans. However, asking students to record or discuss their thinking about their process in a journal, with their peers, or with their teachers (us!) is an easy way to make sure that they’re thinking for themselves rather than using AI, and it can be done with little to no modifications to existing activities.
Something I have been wondering about is not necessarily how my students will be using AI, but how (or IF) I should use AI in my own lesson planning. As I said before, novice teachers often face cognitive overload and tight time constraints when trying to come up with activities for our students. AI has emerged as a possible solution for teachers who want to create lesson plans and materials FAST– but there are lots of ethical concerns, such as intellectual honesty and environmental impact. Where do you stand on teachers using AI to generate class materials? Do you think it’s even possible to use AI to create AI resistant tasks?
Thanks!
I appreciate your perspective! AI is developing so rapidly that these recommendations for “AI resistant” strategies may not hold up for long. For now, they’ve been super helpful as I reimagine writing tasks with teachers. I am also focused on how we can move writing into the classroom. That way, we are not sending this cognitively challenging work home with students to do after a long day when they may be more tempted to use AI to complete work.
In response to your question about using AI to generate lessons, I am an advocate for using AI to elevate our design process. I do not love the idea of throwing a standard or learning objective into an AI-powered education tool and having it spit out the same whole group teacher-led lessons we have been using for years. Instead, I want educators to harness the power of AI to design more equitable and accessible learning experiences in a fraction of the time it would have taken before AI. Time is a luxury educators do not have so the impact that AI can have on our ability to elevate our design work is very exciting.
Catlin
Hi Dr. Tucker, you’re giving me more to think about each time. I am a math teacher and it is incredibly easy for students to use AI in order to get the easy answer. It has been super frustrating when seeing students perform well on homework or math processes but when it comes to assessment time, they have no idea what they are doing. I like the idea of proving their answer, and the personal anecdote of “Where I struggled/how I overcame it”. Even with the proof of a math concept, AI can be useful. This gives me a lot to think about and how I can incorporate it into my classwork.
Yay! I am thrilled these blogs are useful and thought-provoking. The best teachers are lead learners!
I wanted to share my math journal user guide which I developed while coaching a math teacher who wanted to create more accountability around student work and more transparency about their mathematical thinking. If you can use it to combat some of the AI concerns by asking students to surface their thinking in writing, please do!
Take care.
Catlin