Scully and Castro Fernandez challenge free AI training data access

Jess Scully and Raul Castro Fernandez argue that AI companies should compensate creators for open internet data used in model training. Writing for John Menadue’s public policy platform, the authors contend that commercially valuable human-generated content warrants payment despite its public accessibility. They challenge the prevailing industry assumption that openly available web data is inherently free for commercial AI development. The article questions whether current scraping practices constitute fair use when models generate significant revenue from collective human output. This argument aligns with growing global pressure to establish licensing frameworks for generative AI training datasets. The piece does not propose specific compensation mechanisms but frames the issue as an economic rights question. Their position adds intellectual weight to ongoing copyright litigation against major AI laboratories. Industry stakeholders have yet to respond directly to this specific commentary.