Goodhart's Law — named after British economist Charles Goodhart — states that when a measure becomes a target, it ceases to be a good measure. It was written about monetary policy in the 1970s, but it describes something happening right now inside every organization using AI to optimize its communications. You start tracking engagement metrics, so your AI drafts for click rate. You optimize for response time, so the tool trains you toward brevity. The measure becomes the goal, and the thing the measure was pointing at — genuine understanding, real persuasion, actual connection — quietly evacuates the building. The 4th-century Confucian philosopher Xunzi argued that ritual forms (li) were originally designed to channel authentic human feeling outward into shared, legible expression — but that forms emptied of their inner life become something worse than nothing, because they produce the appearance of virtue while corroding its source. What Xunzi saw in hollow ceremony, Goodhart measured in economic data: optimizing the symbol eventually destroys the thing it symbolized. The practical discipline here is not to stop measuring — it's to audit, periodically and deliberately, whether your communication metrics are still pointing at what you actually care about, or whether you've been quietly managing the shadow on the wall instead of the object casting it.

Pick one metric you or your team uses to evaluate communication quality — opens, replies, brevity, tone scores, whatever. What is the actual human outcome it was meant to proxy? Are they still correlated, or has the proxy become the product?

Drawing from Confucian philosophy / Behavioral economics — Xunzi (荀子, Xunzi, ~3rd century BCE) with Charles Goodhart (Goodhart's Law, 'Problems of Monetary Management: The UK Experience', 1975)