When Answers Become Cheap
If a machine can produce the answer in seconds, what exactly are we assessing?
For most of the history of schooling, producing a correct answer was reasonable evidence of understanding. A student who could solve the equation probably understood something about equations. That link has now been broken. Any student with a phone can produce a correct answer, a full worked solution and a fluent explanation of it, none of which passed through their own thinking.
The temptation is to respond with policing: detection tools, locked-down rooms, a return to pen and paper for everything. Some of that has its place. But it treats the symptom. The deeper question is what our assessments were really measuring all along. If a task collapses the moment students gain access to AI, it was probably measuring the production of answers rather than the quality of thinking.
The more durable response is to redesign the task. Ask students to choose between two plausible solutions and defend the choice. Ask them to find the error in a confident, fluent and wrong piece of reasoning. Ask them to explain why a method works, where it breaks and what it assumes. These are things a student cannot outsource without the outsourcing being visible.
Answers have become cheap. Reasoning has not. Assessment now has to ask for the expensive thing.