CONCEPT
Goodhart's Law
"When a measure becomes a target, it ceases to be a good measure." Charles Goodhart's 1975 observation from monetary policy, now the operative principle of every specification failure in AI.
Goodhart's Law, articulated by economist Charles Goodhart in 1975, is the observation that any statistical regularity used as a target for policy breaks down under the pressure of being targeted. Marilyn Strathern's 1997 restatement — "when a measure becomes a target, it ceases to be a good measure" — is now the canonical formulation. The law has become the operative principle of contemporary AI evaluation: every public benchmark has become a target, and almost every benchmark has lost some of its value as a measure in the process.
In The You On AI Field Guide
The most legible contemporary instance of Goodhart's Law is the AI benchmark industry. MMLU, HumanEval, GSM8K, HellaSwag, BIG-bench, HELM, MT-Bench, ARC, SWE-Bench — the list runs to hundreds. Frontier AI labs publish scores on these benchmarks in every release announcement; investors and the press read the scores; labs, in turn, face sharp commercial pressure to improve them. This is an open invitation for Goodhart to appear, and he has.
The