Jun 18 @ 3:00 pm – 4:00 pm –

Organized by AI & Equality by Women at the Table and Globethics, HumRights-Bench is the first scientific tool to measure whether AI models understand and apply human rights law. As large language models increasingly shape decisions affecting rights holders worldwide, Member States urgently need rigorous evaluation tools. Our results reveal consistent model failures — giving governments the concrete data needed to regulate AI responsibly.