Can AI companies legally scrape copyrighted content to train their models? This article examines fair use in the United States, fair dealing in India, and the European Union's text-and-data-mining exceptions, questioning whether public accessibility amounts to consent and exploring licensing-based solutions.