What are model cards and datasheets for datasets?

Both are transparency documents. Model cards (Mitchell et al., 2019) accompany a trained model and summarize its intended use, out-of-scope uses, training data at a high level, evaluation results broken down by relevant subgroups, known limitations, and ethical considerations. Datasheets for datasets (Gebru et al.) document a dataset's motivation, composition, collection process, preprocessing, recommended uses, and maintenance — so downstream users understand provenance and consent. The goal is that anyone reusing a model or dataset can judge fitness for their context rather than discovering limitations in production. Many organizations and regulators now expect this kind of documentation, and it's a low-cost practice that materially improves accountability.