Inverse reinforcement learning (IRL) — repurposed for → Use demonstrations (safe exploration)

Concrete Problems cites inverse reinforcement learning twice, for two structurally different purposes. Under Reward Pretraining, IRL is proposed as a defense against reward hacking: train a fixed reward function from demonstrations before any interaction with the environment, so the agent cannot later manipulate its own reward signal (concrete-problems, §"Reward Pretraining:", p. 10). Under Use Demonstrations, the same technique is repurposed to address a different problem, unsafe exploration: the trajectories are used not to lock down a reward function but to build a baseline policy that lets the agent skip risky trial-and-error learning, or bound further exploration around it (concrete-problems, §"Use Demonstrations:", p. 15). The tool is identical; what changes is which failure mode — hacking versus exploration risk — the resulting policy or reward is meant to guard against.