Goodhart's Law

Diabloto96 (-2)
Ruben Bloom (+227)
Ruben Bloom C-Class on the basis of having some explanation of concept plus some elaboration in conjuction with a really strong posts list.
Ruben Bloom
Ruben Bloom (+10/-10)
Ruben Bloom (+9/-9)
Ruben Bloom (+806/-8)
Ruben Bloom (+559/-528)
Vladimir Nesov (+12) /* See also */
Vladimir Nesov (+14) /* See also */

Goodhart's Law is of particular relevance to AI Alignment. Suppose you have something which is generally a good proxy for "the stuff that humans care about", it would be dangerous to have a powerful AI optimize for the proxy, in accordance with Goodhart's law, the proxy will breakdown.  

In Goodhart Taxonomy, Scott Garrabrant identifiedidentifies four kinds of Goodharting:

In Goodhart Taxonomy,Taxonomy, Scott Garrabrant identified four kinds of Goodharting:

Goodhart's Law states that when a proxy for some value becomes the target of optimization pressure, the proxy will cease to be a good proxy. ConsiderOne form of Goodhart is demonstrated by the Soviet story of a factory graded on how many shoes they produced (a good proxy for productivity) – they soon began producing a higher number of tiny shoes. Useless, but the numbers look good.

Goodhart Taxonomy

In Goodhart Taxonomy, Scott Garrabrant identified four kinds of Goodharting:

  • Regressional Goodhart - When selecting for a proxy measure, you select not only for the true goal, but also for the difference between the proxy and the goal.
  • Causal Goodhart - When there is a non-causal correlation between the proxy and the goal, intervening on the proxy may fail to intervene on the goal.
  • Extremal Goodhart - Worlds in which the proxy takes an extreme value may be very different from the ordinary worlds in which the correlation between the proxy and the goal was observed.
  • Adversarial Goodhart - When you optimize for a proxy, you provide an incentive for adversaries to correlate their goal with your proxy, thus destroying the correlation with your goal.
Load More (10/12)