Most institutions met generative AI with a ban no one can enforce; a credible policy defines acceptable use, redesigns assessment to measure what AI cannot fake, and gives faculty a defensible line between learning aid and misconduct.
The first institutional response to generative AI was, almost everywhere, prohibition. It was understandable and it has largely failed. A ban on tools students carry in their pockets, that produce undetectable text, and that they will use in the workplace is a rule that exists on paper and dissolves in practice. A policy that cannot be enforced is not a policy; it is a posture.
A credible policy starts from an honest premise: the tools are here, students will use them, and the job is to define how — a graded articulation of acceptable use, where AI is a legitimate aid, where it must be disclosed, and where it is misconduct. But policy language alone does not hold, because the deeper problem is assessment design. If an assessment can be aced by AI, it was already measuring the wrong thing. The durable response is to redesign toward what AI cannot fake: in-person reasoning, oral defence, application to novel local contexts, process made visible.
From banning a tool to governing its use, from assessments AI can ace to assessments that measure judgement, from a policy on paper to a practice with process behind it.