How generative AI exposed a flaw in the way we assess learning, and why fixing it requires more humans, not more technology.
