A curated dataset of publicly available resources for AI ethics in media production: Metadata, documentation status, and reuse potential.

Description: This dataset contains structured metadata of 28 publicly available datasets and benchmarks for evaluating ethical dimensions of artificial intelligence (AI) in media production contexts. The resources cover bias detection, authenticity, governance, and flagship ethical datasets released between January 2024 and March 2026. Files included: AI_Ethics_Media_Datasets.xlsx – Main dataset (28 rows, 10 columns) codebook.xlsx – Data dictionary with column descriptions README.txt – Overview, license, and contact information Key fields per resource: Dataset name, year, category, data type, repository, license, documentation status, geographic focus, media domain, and persistent identifier (DOI/arXiv/URL). Data collection:Systematic search of arXiv, Google Scholar, Scopus, Hugging Face, and Zenodo. Inclusion criteria: publicly accessible, explicit ethical focus, relevance to media production (news, advertising, visual communication, film). License: Creative Commons Zero (CC0) 1.0 Universal Related publication: This dataset accompanies a Data in Brief article (DOI to be assigned). Please cite both the article and this dataset.