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.

As generative AI systems increasingly reproduce copyrighted material, traditional infringement rules struggle to determine who should bear responsibility. This article examines whether existing secondary liability standards can adequately address copyright infringement caused by AI hallucinations.