Academy K

Reward Hacking and Specification Gaming

InterviewConceptualreward hackingalignmentspecification

Related module: Alignment fundamentals

Problem statement

Reward hacking occurs when an agent optimizes a proxy reward in ways that violate the designer's intent.

  1. Give three examples of reward hacking in RL systems.
  2. Why is reward hacking sometimes called "Goodhart's Law for AI"?
  3. What approaches can mitigate reward hacking?

Practise this problem

Verify you understand it—not just that you read the answer.

Learn the foundations

The AI / SI track covers this topic and more in depth.