Campbell's Law

Campbell's law states that when a quantitative indicator is used for high-stakes decisions, it distorts the process it is meant to track—people optimize the number, not the underlying goal.

It is a close cousin of Goodhart's law: measures turn into targets and stop reflecting the quality you cared about. Under pressure, shortcuts, teaching-to-the-test behavior and outright fraud become more likely.

Mitigations mirror Goodhart: combine multiple signals, include qualitative review, separate learning metrics from punishment metrics and revisit what each number was supposed to proxy.