Reinforcement learning from human feedback is a machine learning (ML) that incorporates human feedback into the rewards function to help AI models better align with human goals.