Class 7 · CBSE AI · Strand A — Systems Thinking
Goodhart's Law — when AI optimises the wrong thing
'When a measure becomes a target, it stops being a good measure.' Why AI chases the wrong goal. For Class 7.
Class 7 · CBSE AI · Strand A — Systems Thinking
'When a measure becomes a target, it stops being a good measure.' Why AI chases the wrong goal. For Class 7.
Board exam preparation in India
Every student in India knows the difference between 'studying for boards' and 'actually understanding the subject.' The board exam is the proxy metric — it's supposed to measure knowledge, but when it becomes the target, students memorise model answers and forget the concept the day after the exam. The metric rose; the goal didn't.
Call centre targets
A call centre sets a target of 'handle 40 calls per day per agent.' Agents start ending calls early to hit the number — customers' problems remain unsolved. The manager measures productivity; what actually got measured was the speed of getting customers off the phone. Goodhart's Law in every Indian office that has ever had call-time targets.
Every Dhee Learning session for this concept follows three stages. We share the questions Dhee actually asks, so you can hear what a session sounds like.
Stage 1 — Surface
A coaching app rewards its coaches on one number: average test score. Parents were promised real understanding. What do you predict happens to teaching over the next six months?
Rote answer
"Coaches might cheat or only teach exam tricks."
Understood
"The score is a proxy for real understanding. Once coaches are judged on the score, they optimise the number — drilling past papers, rote answer-keys, teaching to the test — while real understanding, which the exam only partly captures, quietly stalls. The proxy rises; the goal it stood for erodes. That is Goodhart's Law: the measure stopped being a good measure once it became the target."
Stage 2 — Reasoning
A hospital rates itself on how many patients each doctor sees per hour, and optimises to push that number up. What is the real goal the hospital should be chasing — and how could 'patients per hour' pull against it?
Follow-up Dhee may use: Can you design a metric that is closer to the real goal and harder to game? What would you measure instead?
Stage 3 — Application
Here are three systems and the metric each one chases. For each, name the gap between the metric and the true goal, and one way the system might 'game' the number: (a) a news app chasing clicks, (b) a fitness app chasing daily step count, (c) a support chatbot chasing 'calls marked resolved'.
Misconception Dhee watches for: Child assumes the metric and the goal stay aligned as long as the designers picked the metric in good faith.
Dhee turns this concept into a short spoken lesson — teaching, listening, and probing — so your child builds the idea themselves.
'When a measure becomes a target, it stops being a good measure.' Why AI chases the wrong goal. For Class 7.
If a metric is improving, the underlying goal must be improving too.
Dhee opens with a question — for example: "A coaching app rewards its coaches on one number: average test score. Parents were promised real understanding. What do you predict happens to teaching over the next six months?" — listens to your child's answer, then probes the reasoning behind it. The session ends when the child can apply the idea to a brand-new situation, not just recall it.