A Teacher Used AI to Fight AI — and Discovered Something Unexpected

AI-assisted cheating has become one of the biggest headaches for teachers around the world. Rather than simply banning AI tools from the classroom, one teacher decided to use the very same "problematic" technology to build a solution. He developed an app designed to generate assignments that are harder to complete using only a chatbot, according to EdSurge.

What an "AI-resistant" assignment looks like

The core idea was simple: if students can type any standard prompt into a chatbot and get a ready-made answer, then assignments need to be rethought. The teacher used AI to design tasks that require personal context, step-by-step reasoning, or references to discussions previously held in class — elements a language model can't easily guess or invent.

In practice, the app was built to generate questions tailored to each group of students, based on recent classroom activities, examples discussed in lessons, or personal experiences. This way, a generic answer produced by ChatGPT would immediately stand out as too vague or out of step with the context.

A discovery more important than beating cheating

Although the original goal was to prevent academic dishonesty, the teacher arrived at a broader conclusion: the process of building "AI-resistant" assignments forced him to think more carefully about what actually makes a good piece of pedagogy. Tasks created this way turned out to be more relevant and more personalized, regardless of whether students used artificial intelligence or not.

The takeaway from the experiment, EdSurge notes, is that "the fight against AI pushed him to fundamentally rethink assignment design." In other words, the technological challenge became an opportunity for real improvement in teaching methods, not just an obstacle to work around.

What this means for teachers everywhere

The example points to a shift in perspective that's becoming increasingly common in education: instead of viewing AI purely as a threat, some teachers are starting to use it as a tool for instructional design. The result isn't just a barrier against cheating, but assignments that are more thoughtful, more personalized, and potentially more effective for learning.