Language data is data made up of human language, including text, speech transcripts, audio, annotations, documents, conversations, or linguistic metadata. It is used to train, evaluate, or operate natural language processing and speech systems.