Data annotation

Data annotation

The labelling of raw data, such as tagging objects in images or rating the quality of a chatbot's replies, so it can be used to train or evaluate an AI model. Data annotation is often outsourced to large, low-paid workforces, chiefly in the Global South, which has become a significant labour and ethics issue in the AI supply chain. (See also: Training data, Reinforcement learning from human feedback)