Artists demand AI firms retrain models on consented data only

Digital artists and critics are publicly demanding that artificial intelligence companies discard existing models trained on unlicensed artwork and retrain systems exclusively using data from creators who provided direct consent. This call, amplified across social media platforms this week, challenges the foundational data practices of major generative AI developers. Proponents argue that current models constitute unauthorized exploitation of creative labor and intellectual property. Industry representatives have previously defended training methodologies as fair use or transformative, though no company has committed to full dataset replacement. The debate intensifies pressure on AI firms to implement verifiable opt-in mechanisms amid ongoing copyright litigation. Legal experts note that such demands exceed current regulatory requirements but reflect growing public sentiment regarding creator compensation. Failure to address these concerns may accelerate legislative action targeting training data transparency and licensing frameworks in key markets.