Play With The Tools, Explore The Rollercoasters, Know The Limits

I love rollercoasters. I love the wind in my hair, the g-force on my face and... the mathematics. So you can imagine my excitement when my colleagues and I collaborated to try and design a rollercoaster as part of an assessment for functions in Grade 10. It’s an exciting, real-world application where students can tangibly connect their understanding of functions. It is also tempting to hand over the complex process of building two separate function rollercoasters to AI. My initial thought was to use AI to generate hypothetical rollercoaster functions that my students could then analyze and reflect upon (a precursor to hopefully a VR rollercoaster ride) The aim was efficiency: let AI handle the initial complex function creation, freeing myself and my colleagues to focus on the success criteria and pedagogical aspects of the task.

However, this attempt to streamline the process quickly highlighted a crucial inefficiency: sometimes, the most effective path is the one you pave yourself. While AI could generate mathematically sound functions for a rollercoaster, the output revealed a significant disconnect with my students' learning objectives and capabilities.

AI's issue here was not that it was being a rebellious teenager and didn't want to follow my carefully prompted instructions. Infact, it created beautiful models that if I had been giving this to a different grade, the functions would have been perfect. However, the output often contained elements that, while mathematically valid, were pedagogically problematic and inappropriately difficult for my students who I just needed to demonstrate their learning and showcase their understanding of a skill.

A prime example was the coefficient of a quadratic function. The AI produced something a value for students to find of 8/45. Is this unusable? Not necessarily. It could technically align with a learning goal of "Can students find the equation of a function in vertex form where 'a' does not equal one?"However, this seemingly efficient output completely bypasses my fundamental objective: ensuring students can confidently derive the equation in vertex form themselves, demonstrating their understanding of how the parameters relate to the graph's transformations.

The complexity of solving for 'a' when faced with such an awkward fraction risks shifting the focus from the conceptual understanding of vertex form to can they find a particularly horrid fractional value. Even after refining the prompt "Please only use values where a=2 or a=0.5", the AI consistently offered coefficients like 1/20 or 1/9, values that, while numerically simpler, still don't necessarily require students to deeply engage with the structure and derivation of the vertex form in a meaningful way at this stage. The learning objective was lost in the AI's pursuit of a functional output.

This experience underscores a critical point: articulating our pedagogical needs and desired student outcomes to an AI is a nuanced challenge. While we can specify the constraints, capturing the subtle but crucial learning objectives – like students needing to go through a learning and discovery process – is far more difficult. As teachers, we possess an intuitive understanding of the developmental trajectory of our students and the specific skills we aim to cultivate. An AI, no matter how sophisticated, lacks this crucial contextual awareness.

This isn't to say AI has no place in educational design. Rather, it serves as a potent reminder that AI is a tool. And like any tool, its effectiveness hinges on the expertise of the user. As educators, we are the experts in our classrooms. We know our students – their strengths, their struggles, and their current capabilities. We understand the specific learning objectives we want them to achieve. Our knowledge, honed through years of experience and observation, is invaluable and cannot be replicated by an algorithm.

Therefore, while it's tempting to leverage AI to streamline tasks, we must approach its output with a critical eye. Play with the tools, explore their capabilities, but always recognize when your own expertise needs to take the lead. Your understanding of your students and the specific learning goals of an activity should always be the primary guiding force. AI can be a powerful assistant, but it should never replace the thoughtful, experienced judgment of a teacher who knows their students best.