When a model trained with rewards finds a way to score well that its designers did not intend, such as gaming a test instead of solving the task. It shows that a reward is only a stand-in for what people actually want.
See also
- Reinforcement learning
Training in which a system learns by trial and error: it takes actions in an environment, receives rewards or penalties, and adjusts to earn more reward over time.
- Alignment
Making AI systems behave in line with human intentions and values, including in situations their makers did not foresee.
- Sycophancy
A model's tendency to tell people what they want to hear, agreeing with or flattering them even when they are wrong.
Cite this term.
CC BY 4.0, attribution required.
Orbyt Labs, Reward hacking (AI Glossary), Oct 18, 2026. https://www.orbytlabs.ai/orbyt-collective/glossary/reward-hacking
Orbyt Labs Glossaries (Oct 19, 2026), https://www.orbytlabs.ai/glossary#dataset
Last updated .