Thousands of Artists, Including Taylor Swift, Found in AI Training Datasets Without Permission

Researchers have identified thousands of artists whose work appears in AI training datasets, from legendary Malaysian filmmaker P. Ramlee to contemporary pop star Taylor Swift, raising significant copyright and consent concerns. The discovery highlights a growing tension between AI developers' data collection practices and artists' intellectual property rights. The findings suggest widespread use of copyrighted material in training generative AI systems without explicit permission or compensation.

Why it matters: This investigation underscores a critical legal and ethical issue for the AI industry—the use of copyrighted creative work in training datasets—that could trigger major regulatory action and litigation affecting how companies can legally source training data.