Ask for a wrong worked example on purpose

Hand-crafting a convincingly wrong solution is slow. AI is excellent at it: ask for a worked example that commits one specific, common error while looking otherwise professional, and you have instant material for a "find the flaw" activity. The pedagogy is well evidenced; explaining the error in an incorrect example deepens conceptual understanding (Booth et al., 2013). AI just removes the production cost.

In practice

Write a worked solution to this problem that a confident Grade 9
student might produce, containing exactly one error: [the
misconception you're targeting]. Make the rest fully correct and
neatly presented. Do not label the error.

• Project it and run "defend or demolish": pairs must either justify the solution or locate the flaw and explain why it's tempting.
• Always check the output yourself first. Models sometimes add a second, accidental error, and you want to know where the planted one is.
• This pairs with "Ask why the wrong answer is tempting" (Understanding): same pedagogy, now with unlimited material.