On August 2, 2026, the EU AI Act starts treating any system that screens, scores, or filters job candidates as high-risk. Annex III, point 4. The text is short, the obligations are not: human oversight before a rejection, audit logs,...