10.7910/DVN/7C6FNO Samuel, Jim Jim Samuel https://orcid.org/0000-0002-7599-3209 Siritha Chidipothu https://orcid.org/0009-0004-1482-1165 Khanna, Tanya Tanya Khanna https://orcid.org/0009-0009-0107-4160 Lakra, Ashish Ashish Lakra https://orcid.org/0009-0006-7581-2240 Vidhi Gala Vidhi Gala Harvard Dataverse 2026 Computer and Information Science Social Sciences Jim Samuel Jim Samuel Tanya Khanna Tanya Khanna Ashish Lakra Ashish Lakra Vidhi Gala Vidhi Gala 2026-06-03 2026-06-05 2026-06-09 1047600713 37088436 38461078 text/comma-separated-values text/tab-separated-values text/tab-separated-values 1.1 Creative Commons CC0 1.0 Universal Public Domain Dedication. The Global Artificial Intelligence News Headlines (GAIN-H) is an open-access public informatics collection of three complementary datasets containing over 2.5 million artificial intelligence-related news headlines gathered from global news sources across multiple languages, countries, and time periods. The repository was created to support interdisciplinary research on how artificial intelligence is represented, framed, and discussed within the public sphere. The collection includes: (1) a metadata-rich corpus with temporal, linguistic, and URL-structural features; (2) a large-scale longitudinal corpus optimized for temporal analysis; and (3) an extended multilingual corpus containing search-term metadata that enables keyword-stratified analysis of AI discourse. Together, these datasets span more than two decades of AI-related news coverage and provide researchers with resources for studying media framing, sentiment, public discourse, AI governance, communication, computational social science, and natural language processing. The repository is intended for researchers, policymakers, educators, journalists, practitioners seeking to examine trends in AI-related media coverage across time, geography, language, and thematic domains. The datasets are released to promote transparency, reproducibility, and evidence-based research on the societal implications of artificial intelligence. The datasets were developed as part of the RAISE (Rethinking AI for Shared Empowerment) initiative at the MPI Program, Bloustein School, Rutgers University, and AIXosphere AI behavioral trends research.