Editing 2899: Goodhart's Law
Warning: You are not logged in. Your IP address will be publicly visible if you make any edits. If you log in or create an account, your edits will be attributed to your username, along with other benefits.
The edit can be undone.
Please check the comparison below to verify that this is what you want to do, and then save the changes below to finish undoing the edit.
Latest revision | Your text | ||
Line 39: | Line 39: | ||
===Additional examples of Goodhart's Law=== | ===Additional examples of Goodhart's Law=== | ||
− | * The classical example of Goodhart's Law is the {{w|Perverse_incentive#The_original_cobra_effect|Cobra Effect}}: anecdotally the British rule in India paid bounties for dead cobras as a pest control effort. | + | * The classical example of Goodhart's Law is the {{w|Perverse_incentive#The_original_cobra_effect|Cobra Effect}}: anecdotally the British rule in India paid bounties for dead cobras as a pest control effort. This worked at first, but soon people began breeding cobras for income. |
− | * | + | * A school's exam results may ''suggest'' how well the school works with its pupils, but may lead to rigidly "teaching to the exams" and lead to less enjoyment and ability of life-long learning, or even flexibility in non-academic activities. |
− | * A hospital measures inpatient ''Length of Stay'' because shorter stays save money and free up beds for | + | * A hospital measures inpatient ''Length of Stay'' because shorter stays save money and also free up beds for any admitted patients waiting in the ER. But if improperly incentivized, doctors may discharge inpatients too early, causing some to "bounce back" to the hospital as a costly readmission. |
− | * A call center measures the number of calls handled per hour | + | * A call center measures the number of calls handled per hour, but poorly decides to overly incentivize this metric to make the workers more productive; that leads workers to cut calls short, frustrating customers. |
* The hypothetical {{w|Instrumental convergence#Paperclip maximizer|Paperclip Maximizer}} concept demonstrates how having a seemingly benign metric as a goal might still result in almost unlimited adverse effects, if unchecked. | * The hypothetical {{w|Instrumental convergence#Paperclip maximizer|Paperclip Maximizer}} concept demonstrates how having a seemingly benign metric as a goal might still result in almost unlimited adverse effects, if unchecked. | ||