Y10W20RC Steelman Then Respond

Steelman Before You Strike

Moderator:

Good afternoon. Today’s debate motion is: ‘Schools should allow the limited use of generative AI tools for homework drafting and study support.’ This is not a debate about replacing thinking with software. It is a debate about whether a school can create a careful policy that allows some use while still protecting learning, fairness and academic integrity. Our speakers have agreed to follow one extra rule. Before each major disagreement, they must give a ‘steelman’ restatement of the other side. That means presenting the strongest fair version of the opposing view before responding to it.

Speaker 1, for the motion:

Thank you. I support limited use because students already work in a world full of digital tools, and schools should teach wise use rather than pretend those tools do not exist. A limited policy could allow students to use AI for brainstorming, planning, clarifying explanations and checking whether their ideas are organised, while still banning submission of machine-generated final answers as if they were fully original work. My case rests on three reasons: preparation, access and transparency.

First, preparation. Students will enter workplaces and further study environments where digital assistance is common. If school only teaches prohibition, students may leave without the judgement needed to use such tools responsibly. Second, access. Some students get help at home from tutors, older siblings or highly educated parents. Others do not. A carefully restricted tool can offer a baseline level of support for planning and explanation. Third, transparency. At the moment, many students use these tools secretly or unevenly. A limited policy could require simple disclosure, such as a note stating whether AI was used for idea generation, feedback or structure. That would create clearer norms than silence does.

I am not arguing that convenience is more important than learning. I am arguing that thoughtful policy can distinguish between support and substitution. Schools already accept calculators in some maths tasks, spellcheck in many writing contexts and online databases for research. The educational question is not ‘Is a tool powerful?’ but ‘What kind of thinking should students still do themselves?’ If that boundary is clearly defined, limited use can strengthen rather than weaken learning.

Speaker 2, against the motion:

Thank you. I oppose the motion because even limited permission creates confusion about authorship, weakens practice in essential thinking and shifts school culture towards assisted performance instead of independent effort. My position is not anti-technology. It is pro-clarity. When students are still learning how to generate ideas, shape arguments and find their own wording, outsourcing even the early stages can quietly reduce the mental work that builds skill.

My first reason is cognitive development. Drafting is not just a final step after thinking; drafting is part of thinking. When students struggle to begin, refine a sentence or choose between examples, that effort is not wasted. It is the process through which judgement grows. My second reason is enforceability. A ‘limited use’ policy sounds balanced, but in practice it becomes difficult to police. Where exactly does brainstorming end and ghostwriting begin? How will teachers distinguish between a student lightly assisted by software and one heavily shaped by it? My third reason is trust. If every polished paragraph may or may not have invisible assistance behind it, confidence in assessment weakens.

Supporters sometimes describe disclosure as the solution, but disclosure depends on honesty, and honesty becomes harder to sustain when the permitted and forbidden uses sit close together. A policy should not only sound reasonable on paper. It should also be workable in real classrooms, across different teachers, subjects and year levels. In my view, a school serves students better by teaching strong human drafting habits first and reserving AI discussion for teacher-led demonstrations rather than student homework use.

Steelman turn

Moderator:

Thank you. We will now move to the first steelman turn. Speaker 1, before replying, please restate the strongest version of Speaker 2’s position.

Speaker 1:

The strongest version of my opponent’s case is this: writing and planning are not merely ways of displaying thought after it happens. They are part of how thought becomes sharper. If schools permit AI too early or too loosely, students may skip the very struggle that helps them develop voice, reasoning and endurance. In addition, even a well-meant policy could be difficult to monitor consistently, which might create unfairness between students and uncertainty for teachers. That is a serious concern, not a fearful one.

My response is that this concern describes a bad policy, not every possible policy. I agree that students must still do the core cognitive work. That is why I support limited use, not open use. For example, schools could ban AI for final phrasing and evidence selection while allowing it for question clarification, brainstorm prompts and revision planning. Teachers already manage boundaries around collaboration, online research and drafting support. Complexity is not a reason to avoid policy altogether. Sometimes it is a reason to write policy more carefully.

Steelman turn

Moderator:

Speaker 2, please restate the strongest version of Speaker 1’s position before responding.

Speaker 2:

The strongest version of the case for the motion is this: students already live among intelligent digital tools, and pretending otherwise may produce secrecy rather than integrity. A carefully limited policy might bring hidden behaviour into the open, allow students to learn responsible habits and offer some support to those who do not have much academic help outside school. In that sense, the argument is not about laziness. It is about realism, equity and explicit guidance.

My response is that realism does not automatically require permission. Schools also respond to reality by deciding which pressures to slow down, not only which tools to absorb. The motion assumes that transparency becomes easier once use is allowed, but that may not happen. Students may disclose some use and hide other use. Teachers may spend more time investigating process instead of teaching content. Most importantly, a school’s first job is not to mirror every outside tool immediately. It is to protect the conditions under which students can build original competence.

Rebuttal cycle

Speaker 1:

I would like to address the idea of competence more directly. My opponent is right that original competence matters. But competence in modern writing increasingly includes evaluating suggestions, rejecting weak phrasing and deciding when a tool has crossed a line. Those are not trivial skills. They require discernment, which means careful judgement between stronger and weaker options. If a student uses AI to generate three possible structures for an essay and then selects, revises and justifies one of them, that student may still be doing serious thinking. The key is not whether assistance appeared at all. The key is whether the student remained intellectually accountable for the work.

Speaker 2:

That sounds persuasive, but it risks turning every hidden shortcut into a lesson in judgement. A student who receives three possible structures has already been spared the uncertainty of starting from nothing. That matters. The blank page is difficult, yet learning how to move from uncertainty to structure is part of the educational task. If we make support too available at the moment of difficulty, we may produce smoother homework and weaker thinkers. The policy may look compassionate while slowly reducing resilience.

Speaker 1:

I do not think all struggle is equally valuable. Some struggle teaches. Some struggle merely delays. If a student is stuck because they cannot see how to break a large task into parts, a limited prompt that models possible steps may free them to do the deeper work of argument and evidence. We should be careful not to romanticise frustration. Schools regularly provide scaffolds, exemplars and planning frameworks. AI can be treated as one more scaffold if the expectations are explicit.

Speaker 2:

The difference is that a scaffold designed by a teacher is visible, shared and tied to the lesson’s purpose. An AI system is responsive, private and variable. Two students may receive entirely different kinds of help, and a teacher may never know how much shaping occurred before the draft arrived. That variability weakens comparability. In assessment, comparability matters. If schools blur the origin of student work, even with good intentions, they risk sending mixed messages about independence.

Moderator:

Before we move to closing statements, one brief clarification question for both speakers. What would fairness look like under your preferred policy?

Speaker 1:

Under my preferred policy, fairness would mean explicit limits, required disclosure, teacher-designed tasks that still demand personal reasoning, and consequences for crossing the line from support into substitution. Students would know the boundary in advance, and teachers would discuss examples openly.

Speaker 2:

Under my preferred policy, fairness would mean all students completing homework without AI-generated drafting support, while still receiving teacher guidance, class scaffolds and ordinary non-generative tools. The boundary would be simpler, clearer and more consistent across classrooms.

Closing statements

Speaker 1:

This debate is not about surrendering education to software. It is about whether schools can teach responsible tool use without abandoning human thinking. I have argued that they can. A limited policy could prepare students for the real world, reduce secrecy and support students who need a starting point, while still protecting the core expectation that ideas, judgement and final accountability remain human. If we want students to disagree thoughtfully in a changing world, we should teach them how to use powerful tools with discipline, not only how to avoid them.

Speaker 2:

I have argued that the strongest case against the motion is not panic but priority. Students need protected space to practise the difficult mental work of beginning, shaping and refining thought for themselves. Once a school allows AI into that process, even in limited ways, the line between assistance and authorship becomes harder to hold. Respectful disagreement begins with representing the other side fairly, but it also requires deciding what should be preserved. In this case, I believe schools should preserve the habit of independent drafting before they permit machine assistance.