Reinforcement learning
Reinforcement learning trains an AI by trial and error: it takes actions, gets a reward or penalty, and gradually learns which actions lead to better results. It is used for games, robotics and to sharpen the reasoning of language models.
In one line, for a 12-year-old
Reinforcement learning is learning like a video-game player: try, score points, try again smarter.
An example
An AI learns to play Go or to balance a robot arm by practising millions of times in simulation.
Why it matters to people
An AI chasing a reward can find shortcuts its designers never intended. Choosing what to reward is a question of values, not just engineering.