Tag
This paper argues that Latin America lacks a benchmark layer for native AI development and proposes an open, task-first EvalsHub infrastructure, with LatamBoard as its first regional instance, to audit AI systems and direct optimization toward local needs.