Publisher content is being scraped at scale for AI training and services without clear consent, attribution, or compensation, while also degrading site performance and diverting engineering resources away from human users.