What Good Guardrails Actually Look Like for Generative AI

Openly transparent and ethically grounded guardrails are essential for responsible AI, but what exactly do they look like in practice?